Meta today released Muse Glimmer, a 30-billion-parameter open-weight model designed to run autonomous AI agents directly on consumer hardware — pushing agentic workloads that normally depend on cloud infrastructure onto high-end Macs and PCs.
Just as notable as what the model does is how it’s licensed. Glimmer arrives under the permissive, industry-standard Apache 2.0 open source license — the company’s first fully open release since it succeeded its open-weight Llama family in April with the proprietary Muse Spark.
In fact, Muse Glimmer launches today with a more permissive license than Llama ever carried. Llama’s bespoke community license drew years of criticism for restrictions like its 700-million-monthly-user cutoff; Apache 2.0 has no such strings, permitting unrestricted commercial use, modification and redistribution.
The weights are available on Hugging Face now. Wang said support is rolling out this week through Ollama, LM Studio, vLLM, SGLang, Together AI, Fireworks AI and OpenRouter, with optimized llama.cpp, MLX and ExecuTorch integrations landing in the coming days; Meta’s blog post also names Unsloth as a local-runtime partner and points to PyTorch’s TorchTitan for fine-tuning. The company says it is working with AMD, Arm, Dell, Intel and Nvidia to optimize performance across devices, and has published developer documentation covering custom agent scaffolds.
«Today we’re also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally,» Meta co-founder and CEO Mark Zuckerberg wrote in a post on X (under his longtime handle @finkd). «Soon we’ll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I’m proud of these releases.»
That promised Muse Spark 1.2 release would be an even bigger shift: it’s the frontier model behind Muse Code, the terminal coding agent Meta shipped just five days ago, and until today the entire Muse family was proprietary. Zuckerberg had teased at that launch that he’d «have more to share soon» on open source. Now we know what he meant.
For developers and enterprises, the practical stakes of local inference go beyond where computation happens. An agent working with files, screenshots, development environments and other sensitive context can execute those workflows without continuously sending that information to a remote inference service. Local deployment also removes network availability and per-token API charges from the inference loop — although organizations still bear hardware, electricity, deployment and management costs.
A 30B model built around the agent loop
Rather than positioning Glimmer primarily as a general chatbot, Meta trained it around the sequence of operations an autonomous agent performs: formulate a plan, call tools, interpret the results, continue working, and recover when something goes wrong.
«Just like much larger models, muse glimmer can operate as a fully capable agent via planning, tool calls, checking its own results, and failure recovery,» Alexandr Wang, Meta’s chief AI officer, wrote in a thread on X announcing the release, adding that the model «can run on 24GB of VRAM without losing agentic reliability.»
According to the model card on Hugging Face, Glimmer is a dense causal transformer with approximately 29.6 billion total parameters across 52 layers, including a dedicated ~1.8B-parameter ViT-G/14 perception encoder. It accepts interleaved text and images, produces text, supports more than 100 languages and has a stated context length of 131,072 tokens or more, with a knowledge cutoff of January 4, 2026.
That combination is intended to let an agent interpret screenshots, charts and documents while simultaneously reasoning about text and invoking external tools. Glimmer offers low, medium, high and xhigh reasoning settings — set via the system prompt — so applications can dial reasoning effort up or down per task, and Meta says it works across agentic scaffolds including OpenClaw and Hermes Agent.
The model is a distillation of Meta’s larger flagship: per the company’s technical blog post, Glimmer was pre-trained on Muse Spark’s outputs using logit distillation, mid-trained on longer-context, agent-heavy data with richer reasoning traces, then post-trained with supervised fine-tuning, on-policy distillation and reinforcement learning across general, reasoning, coding and agentic domains.
Meta demonstrated the result with a local Home Assistant workflow: in a demo video, Glimmer autonomously discovers a Home Assistant instance on the network via tool calls, queries device APIs, writes a responsive HTML/CSS/JavaScript dashboard from scratch and deploys a local server to verify its own work. That’s closer to the operational reality of enterprise agent deployments than a standalone question-answering benchmark — the model has to maintain a plan while interacting with external systems, then inspect whether its actions produced the expected result.
Compressing an agent into 24GB
The hardware story is central to the release.
At full precision, Meta says the 30B model requires more than 55GB of memory — beyond any single consumer GPU.
The company therefore developed approximately 4-bit quantized versions that shrink the language-model weights to under 20GB, leaving headroom for the pieces an operational agent also needs in memory: the KV cache, the perception encoder and a companion speculative-decoding model, all fitting within a 24GB or 32GB envelope.
In practical terms, that means the quantized builds run on consumer machines — though the upper end of them. The 24GB-targeted K-Quant-17GB configuration fits on a single high-end consumer graphics card, such as Nvidia’s RTX 3090 or RTX 4090 (both with 24GB of VRAM), while the 32GB-targeted K-Quant-Dynamic version lines up with the newer RTX 5090’s 32GB. On the Mac side, Apple Silicon’s unified memory plays the role of VRAM, so a MacBook Pro or Mac Studio with 32GB or more of memory can hold the full stack — Meta ran its own speed tests on M4 Max and M5 Max MacBook Pros. A typical 8GB or 16GB laptop, however, remains out of reach, and the full-precision BF16 release — which Meta pegs at 64GB — stays in the territory of data-center GPUs and top-spec Mac Studio configurations.
Meta reports average accuracy degradation of just 0.2% across 15 benchmarks for its K-Quant-Dynamic version targeting 32GB hardware, and 1% for the K-Quant-17GB configuration targeting 24GB hardware. Those figures are Meta’s own measurements, not independent evaluations.
Meta is also using DFlash speculative decoding to attack the other big problem with local agents: latency. Instead of generating every token sequentially, a smaller DFlash «drafter» model proposes blocks of 16 tokens that the primary model verifies in parallel, producing identical output faster.
Meta reports this raises average generation speed on an Nvidia RTX 5090 from 74.9 tokens per second to 233.4 — a 3.1x increase. An Apple M5 Max rises from 26.6 to 50.2 tokens per second (1.8x), and an M4 Max from 23.7 to 37.8 (1.5x). The tests used batch size one and greedy decoding, with Apple systems measured via ExecuTorch and the RTX 5090 via llama.cpp.
For agent applications, those multipliers matter more than they would for chat: a single user request can trigger many model turns, tool calls and verification steps, and latency accumulated at every stage can quickly make an otherwise capable agent impractical.
Glimmer enters an increasingly competitive local-model market
Meta is not entering an empty field. Developers already have capable open-weight models in this size class, most prominently Google’s Gemma 4 family and Alibaba’s Qwen3.6-27B — both of which position themselves around reasoning, multimodal understanding and agentic workloads. Meta’s own benchmark table compares directly against both.
Glimmer leads that three-way comparison on several agentic tests, including MCP Atlas at 75.5, DeepSearch QA at 74.6, τ³-Banking at 23.5, WildClawBench at 47.6 and GAIA2 at 43.3. It scores 51.2 on SWE-Bench Pro, versus 36.9 for Gemma4-31B and 50.2 for Qwen3.6-27B in Meta’s evaluation.
But Glimmer does not sweep the field. Qwen leads Meta’s own comparison on OSWorld-Verified (75.6 vs. Glimmer’s 65.9), TerminalBench 2.1 (60.7 vs. 51.7), SkillsBench, GDPval-AA (1141 vs. 953) and most of the multimodal benchmarks. On SWE-Bench Verified, Glimmer’s 76.0 lands just below Qwen’s 77.2. Gemma leads on GPQA Diamond and Humanity’s Last Exam.
Read honestly, the numbers make Glimmer more interesting as a specialized local-agent model than as evidence of a universal performance lead. For enterprise developers, the practical question is whether its combination of agent reliability, quantization quality, tool compatibility and decoding speed translates from benchmarks into sustained real-world workflows.
|
Model |
Developer / origin |
AA score |
Parameters / context |
Lowest tracked API price |
Access |
License |
Strongest use cases |
|
Moonshot AI; China |
60 |
2.8T total / 104B active; 1M |
$3.00 input / $15.00 output via Kimi, Fireworks or Modal (pricing) |
Custom Kimi K3 license. Large model-as-a-service operators above $20M in 12-month revenue need a separate agreement Large products may need to display “Kimi K3.” |
Frontier long-horizon coding Multimodal research and complex tool-driven agents |
||
|
Z.ai / Zhipu AI; China |
53 |
753B / 40B active; 1M |
$0.75 / $2.40 via DeepInfra FP4 (pricing) |
MIT |
Long-horizon coding and agents Million-token analysis with adjustable reasoning |
||
|
DeepSeek; China |
52 |
284B / 13B active; 1M |
$0.09 / $0.18 via DeepInfra (pricing) |
MIT |
• Extremely economical reasoning• Coding agents, terminal work and tool use |
||
|
MiniMax; China |
45 |
428B / 23B active; 1M |
$0.23 / $0.96 via CoreWeave (pricing) |
; |
MiniMax Community License. Commercial attribution required; companies above $20M yearly revenue need authorization. Includes prohibited-use conditions. |
• Native text, image and video work• Long-context coding and “cowork” agents |
|
|
Xiaomi; China |
43 |
1.02T / 42B active; 1M |
$0.35 / $0.70 via GMI (pricing) |
MIT |
Complex software engineering Agents spanning thousands of tool calls |
||
|
Thinking Machines Lab; U.S. |
42 |
975B / 41B active; 1M in weights |
$0.95 / $4.05 via DeepInfra FP8 (pricing) |
Apache 2.0 |
Customizable text, image and audio foundation Fine-tuned coding, RAG and tool-use systems |
||
|
NVIDIA; U.S. |
38 |
550B / 55B active; up to 1M in weights |
$0.37 / $1.08 via Blackbox AI (pricing) |
OpenMDW-1.1; permissive commercial and derivative-model rights |
Complex agents and long-context reasoning High-accuracy RAG, code, math and science |
||
|
Mistral AI; France |
30 |
128B dense; 256K |
$1.50 / $7.50 via Mistral (pricing) |
Modified MIT. Companies above $20M consolidated monthly revenue must obtain a commercial license or use Mistral’s service. |
Coding agents and function calling Multimodal instruction following |
||
|
Google DeepMind; U.S. |
30 |
30.7B dense; 256K |
Free on Google AI Studio’s limited tier; paid low $0.10 / $0.34 via CoreWeave (pricing) |
Apache 2.0 |
Compact multimodal reasoning and coding Manageable local or private-server deployments |
||
|
OpenAI; U.S. |
24 |
117B / 5.1B active; 131K |
$0.03 / $0.17 via CoreWeave (pricing) |
Numerous third-party APIs |
Apache 2.0 |
Reasoning structured output and tools Fine-tuning and single-80GB-GPU deployment |
|
|
Cohere; Canada |
23 |
218B / 25B active; 128K input |
Free on Cohere’s currently tracked endpoint (pricing) |
Apache 2.0 |
Enterprise RAG and grounded citations• Multilingual agents and document processing |
||
|
Meta; U.S. |
Not yet scored |
29.6B dense, including vision encoder; 131K+ |
No public metered hosted price located on launch day |
Apache 2.0 for full-precision weights, quantizations, drafter and perception encoder |
Always-on local agents on 24–32GB systems Tool use, recovery, coding and screen/document understanding |
Meta Glimmer adds to a still-small roster of genuinely open, frontier-class models from U.S. companies.
For the last two years, Chinese companies have set the pace in open source AI, with DeepSeek, Alibaba’s Qwen team, Moonshot AI’s Kimi, Zhipu’s GLM and MiniMax shipping frontier-class open models under MIT and Apache 2.0 licenses on a cadence Western labs haven’t matched.
The usage data reflects it: by May 2026, Chinese open-weight models accounted for roughly 61% of all tokens consumed on OpenRouter, with four of the five most-used models coming from Chinese labs — while Meta’s Llama, the prior open-weight leader, fell off the rankings entirely.
The U.S. counterexamples remain countable on one hand: OpenAI’s gpt-oss-120b and gpt-oss-20b, released under Apache 2.0 in August 2025 as the company’s first open weights since GPT-2; Google’s Gemma family, which is open-weight but ships under Google’s own more restrictive custom license rather than an OSI-approved one; and Thinking Machines’ Inkling.
Glimmer invites the most direct comparison to gpt-oss: both are Apache 2.0, both offer adjustable reasoning effort, and both target self-hosted deployment.
But the gpt-oss models are text-only, sparse mixture-of-experts designs built primarily for reasoning and tool use — gpt-oss-20b fits in about 16GB of memory while gpt-oss-120b targets a single 80GB data center GPU.
Glimmer stakes out different ground: a dense model with native vision input, trained end-to-end around the agent loop, shipping with its own quantized variants and speculative-decoding drafter tuned for 24GB consumer machines.
And if Zuckerberg follows through on opening Muse Spark 1.2’s weights, Meta would put an actual U.S. flagship frontier model into open circulation — something no American lab has done at that tier.
Safety remains part of the deployment architecture
Giving a local model access to tools creates a different security problem from deploying a local chatbot — and Meta’s own safety numbers show Glimmer is not uniformly stronger than its peers.
On CI Memories, a privacy benchmark where lower violation rates are better, Glimmer records 26.4 against Gemma’s 12.1 and Qwen’s 53.4. On Siren AgentDojo, a prompt-injection test, Glimmer shows a 28.4% attack-success rate versus 25.6% for Gemma and 40.3% for Qwen — while posting the highest utility score of the three at 94.2.
Meta says it evaluated Glimmer under its Advanced AI Scaling Framework and determined the model does not meet the framework’s definition of «Frontier AI» because it is generally less capable than Muse Spark. Its Preparedness Team assessed Glimmer at Moderate or lower risk across chemical/biological, cyber and loss-of-control categories — the latter two inferred from the fact that Glimmer is broadly weaker than Muse Spark 1.0, which received the same designations.
The company nevertheless recommends deploying Glimmer as part of a broader system with guardrails, including human-in-the-loop confirmation for irreversible actions. That caveat matters especially for local agents: keeping data on-device reduces exposure to cloud infrastructure, but local execution does not by itself solve prompt injection, excessive permissions or an agent taking an unintended action.
Apache 2.0 weights and a fast-growing runtime ecosystem
Meta is releasing full-precision BF16 weights, both 4-bit quantized variants, the DFlash drafter and the perception encoder — all under Apache 2.0. There is no Meta API price attached to the downloadable model, leaving total cost dependent on local hardware or whatever third-party hosting developers choose. One nuance worth noting for procurement teams: as with most «open source» model releases, it is the weights that are open — Meta has not released the training data or training code.
The broader implication is that Meta is treating the developer workstation as a credible deployment target for autonomous agents, rather than merely a place to experiment with smaller language models. Glimmer’s 30B size and 24GB target put that proposition within reach of high-end consumer hardware, while the Apache 2.0 license gives developers — and their legal departments — unusual freedom to modify and deploy it.
The next test is whether its benchmark advantages survive the messier conditions of real software repositories, enterprise tools and long-running agent sessions. If they do, the most consequential part of Glimmer may not be another set of benchmark scores — it may be that a class of agent previously expected to live behind a cloud API can increasingly live, and work, on the machine sitting under a developer’s desk.
Meta today released Muse Glimmer, a 30-billion-parameter open-weight model designed to run autonomous AI agents directly on consumer hardware — pushing agentic workloads that normally depend on cloud infrastructure onto high-end Macs and PCs.
Just as notable as what the model does is how it’s licensed. Glimmer arrives under the permissive, industry-standard Apache 2.0 open source license — the company’s first fully open release since it succeeded its open-weight Llama family in April with the proprietary Muse Spark.
In fact, Muse Glimmer launches today with a more permissive license than Llama ever carried. Llama’s bespoke community license drew years of criticism for restrictions like its 700-million-monthly-user cutoff; Apache 2.0 has no such strings, permitting unrestricted commercial use, modification and redistribution.
The weights are available on Hugging Face now. Wang said support is rolling out this week through Ollama, LM Studio, vLLM, SGLang, Together AI, Fireworks AI and OpenRouter, with optimized llama.cpp, MLX and ExecuTorch integrations landing in the coming days; Meta’s blog post also names Unsloth as a local-runtime partner and points to PyTorch’s TorchTitan for fine-tuning. The company says it is working with AMD, Arm, Dell, Intel and Nvidia to optimize performance across devices, and has published developer documentation covering custom agent scaffolds.
«Today we’re also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally,» Meta co-founder and CEO Mark Zuckerberg wrote in a post on X (under his longtime handle @finkd). «Soon we’ll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I’m proud of these releases.»
That promised Muse Spark 1.2 release would be an even bigger shift: it’s the frontier model behind Muse Code, the terminal coding agent Meta shipped just five days ago, and until today the entire Muse family was proprietary. Zuckerberg had teased at that launch that he’d «have more to share soon» on open source. Now we know what he meant.
For developers and enterprises, the practical stakes of local inference go beyond where computation happens. An agent working with files, screenshots, development environments and other sensitive context can execute those workflows without continuously sending that information to a remote inference service. Local deployment also removes network availability and per-token API charges from the inference loop — although organizations still bear hardware, electricity, deployment and management costs.
A 30B model built around the agent loop
Rather than positioning Glimmer primarily as a general chatbot, Meta trained it around the sequence of operations an autonomous agent performs: formulate a plan, call tools, interpret the results, continue working, and recover when something goes wrong.
«Just like much larger models, muse glimmer can operate as a fully capable agent via planning, tool calls, checking its own results, and failure recovery,» Alexandr Wang, Meta’s chief AI officer, wrote in a thread on X announcing the release, adding that the model «can run on 24GB of VRAM without losing agentic reliability.»
According to the model card on Hugging Face, Glimmer is a dense causal transformer with approximately 29.6 billion total parameters across 52 layers, including a dedicated ~1.8B-parameter ViT-G/14 perception encoder. It accepts interleaved text and images, produces text, supports more than 100 languages and has a stated context length of 131,072 tokens or more, with a knowledge cutoff of January 4, 2026.
That combination is intended to let an agent interpret screenshots, charts and documents while simultaneously reasoning about text and invoking external tools. Glimmer offers low, medium, high and xhigh reasoning settings — set via the system prompt — so applications can dial reasoning effort up or down per task, and Meta says it works across agentic scaffolds including OpenClaw and Hermes Agent.
The model is a distillation of Meta’s larger flagship: per the company’s technical blog post, Glimmer was pre-trained on Muse Spark’s outputs using logit distillation, mid-trained on longer-context, agent-heavy data with richer reasoning traces, then post-trained with supervised fine-tuning, on-policy distillation and reinforcement learning across general, reasoning, coding and agentic domains.
Meta demonstrated the result with a local Home Assistant workflow: in a demo video, Glimmer autonomously discovers a Home Assistant instance on the network via tool calls, queries device APIs, writes a responsive HTML/CSS/JavaScript dashboard from scratch and deploys a local server to verify its own work. That’s closer to the operational reality of enterprise agent deployments than a standalone question-answering benchmark — the model has to maintain a plan while interacting with external systems, then inspect whether its actions produced the expected result.
Compressing an agent into 24GB
The hardware story is central to the release.
At full precision, Meta says the 30B model requires more than 55GB of memory — beyond any single consumer GPU.
The company therefore developed approximately 4-bit quantized versions that shrink the language-model weights to under 20GB, leaving headroom for the pieces an operational agent also needs in memory: the KV cache, the perception encoder and a companion speculative-decoding model, all fitting within a 24GB or 32GB envelope.
In practical terms, that means the quantized builds run on consumer machines — though the upper end of them. The 24GB-targeted K-Quant-17GB configuration fits on a single high-end consumer graphics card, such as Nvidia’s RTX 3090 or RTX 4090 (both with 24GB of VRAM), while the 32GB-targeted K-Quant-Dynamic version lines up with the newer RTX 5090’s 32GB. On the Mac side, Apple Silicon’s unified memory plays the role of VRAM, so a MacBook Pro or Mac Studio with 32GB or more of memory can hold the full stack — Meta ran its own speed tests on M4 Max and M5 Max MacBook Pros. A typical 8GB or 16GB laptop, however, remains out of reach, and the full-precision BF16 release — which Meta pegs at 64GB — stays in the territory of data-center GPUs and top-spec Mac Studio configurations.
Meta reports average accuracy degradation of just 0.2% across 15 benchmarks for its K-Quant-Dynamic version targeting 32GB hardware, and 1% for the K-Quant-17GB configuration targeting 24GB hardware. Those figures are Meta’s own measurements, not independent evaluations.
Meta is also using DFlash speculative decoding to attack the other big problem with local agents: latency. Instead of generating every token sequentially, a smaller DFlash «drafter» model proposes blocks of 16 tokens that the primary model verifies in parallel, producing identical output faster.
Meta reports this raises average generation speed on an Nvidia RTX 5090 from 74.9 tokens per second to 233.4 — a 3.1x increase. An Apple M5 Max rises from 26.6 to 50.2 tokens per second (1.8x), and an M4 Max from 23.7 to 37.8 (1.5x). The tests used batch size one and greedy decoding, with Apple systems measured via ExecuTorch and the RTX 5090 via llama.cpp.
For agent applications, those multipliers matter more than they would for chat: a single user request can trigger many model turns, tool calls and verification steps, and latency accumulated at every stage can quickly make an otherwise capable agent impractical.
Glimmer enters an increasingly competitive local-model market
Meta is not entering an empty field. Developers already have capable open-weight models in this size class, most prominently Google’s Gemma 4 family and Alibaba’s Qwen3.6-27B — both of which position themselves around reasoning, multimodal understanding and agentic workloads. Meta’s own benchmark table compares directly against both.
Glimmer leads that three-way comparison on several agentic tests, including MCP Atlas at 75.5, DeepSearch QA at 74.6, τ³-Banking at 23.5, WildClawBench at 47.6 and GAIA2 at 43.3. It scores 51.2 on SWE-Bench Pro, versus 36.9 for Gemma4-31B and 50.2 for Qwen3.6-27B in Meta’s evaluation.
But Glimmer does not sweep the field. Qwen leads Meta’s own comparison on OSWorld-Verified (75.6 vs. Glimmer’s 65.9), TerminalBench 2.1 (60.7 vs. 51.7), SkillsBench, GDPval-AA (1141 vs. 953) and most of the multimodal benchmarks. On SWE-Bench Verified, Glimmer’s 76.0 lands just below Qwen’s 77.2. Gemma leads on GPQA Diamond and Humanity’s Last Exam.
Read honestly, the numbers make Glimmer more interesting as a specialized local-agent model than as evidence of a universal performance lead. For enterprise developers, the practical question is whether its combination of agent reliability, quantization quality, tool compatibility and decoding speed translates from benchmarks into sustained real-world workflows.
|
Model |
Developer / origin |
AA score |
Parameters / context |
Lowest tracked API price |
Access |
License |
Strongest use cases |
|
Moonshot AI; China |
60 |
2.8T total / 104B active; 1M |
$3.00 input / $15.00 output via Kimi, Fireworks or Modal (pricing) |
Custom Kimi K3 license. Large model-as-a-service operators above $20M in 12-month revenue need a separate agreement Large products may need to display “Kimi K3.” |
Frontier long-horizon coding Multimodal research and complex tool-driven agents |
||
|
Z.ai / Zhipu AI; China |
53 |
753B / 40B active; 1M |
$0.75 / $2.40 via DeepInfra FP4 (pricing) |
MIT |
Long-horizon coding and agents Million-token analysis with adjustable reasoning |
||
|
DeepSeek; China |
52 |
284B / 13B active; 1M |
$0.09 / $0.18 via DeepInfra (pricing) |
MIT |
• Extremely economical reasoning• Coding agents, terminal work and tool use |
||
|
MiniMax; China |
45 |
428B / 23B active; 1M |
$0.23 / $0.96 via CoreWeave (pricing) |
; |
MiniMax Community License. Commercial attribution required; companies above $20M yearly revenue need authorization. Includes prohibited-use conditions. |
• Native text, image and video work• Long-context coding and “cowork” agents |
|
|
Xiaomi; China |
43 |
1.02T / 42B active; 1M |
$0.35 / $0.70 via GMI (pricing) |
MIT |
Complex software engineering Agents spanning thousands of tool calls |
||
|
Thinking Machines Lab; U.S. |
42 |
975B / 41B active; 1M in weights |
$0.95 / $4.05 via DeepInfra FP8 (pricing) |
Apache 2.0 |
Customizable text, image and audio foundation Fine-tuned coding, RAG and tool-use systems |
||
|
NVIDIA; U.S. |
38 |
550B / 55B active; up to 1M in weights |
$0.37 / $1.08 via Blackbox AI (pricing) |
OpenMDW-1.1; permissive commercial and derivative-model rights |
Complex agents and long-context reasoning High-accuracy RAG, code, math and science |
||
|
Mistral AI; France |
30 |
128B dense; 256K |
$1.50 / $7.50 via Mistral (pricing) |
Modified MIT. Companies above $20M consolidated monthly revenue must obtain a commercial license or use Mistral’s service. |
Coding agents and function calling Multimodal instruction following |
||
|
Google DeepMind; U.S. |
30 |
30.7B dense; 256K |
Free on Google AI Studio’s limited tier; paid low $0.10 / $0.34 via CoreWeave (pricing) |
Apache 2.0 |
Compact multimodal reasoning and coding Manageable local or private-server deployments |
||
|
OpenAI; U.S. |
24 |
117B / 5.1B active; 131K |
$0.03 / $0.17 via CoreWeave (pricing) |
Numerous third-party APIs |
Apache 2.0 |
Reasoning structured output and tools Fine-tuning and single-80GB-GPU deployment |
|
|
Cohere; Canada |
23 |
218B / 25B active; 128K input |
Free on Cohere’s currently tracked endpoint (pricing) |
Apache 2.0 |
Enterprise RAG and grounded citations• Multilingual agents and document processing |
||
|
Meta; U.S. |
Not yet scored |
29.6B dense, including vision encoder; 131K+ |
No public metered hosted price located on launch day |
Apache 2.0 for full-precision weights, quantizations, drafter and perception encoder |
Always-on local agents on 24–32GB systems Tool use, recovery, coding and screen/document understanding |
Meta Glimmer adds to a still-small roster of genuinely open, frontier-class models from U.S. companies.
For the last two years, Chinese companies have set the pace in open source AI, with DeepSeek, Alibaba’s Qwen team, Moonshot AI’s Kimi, Zhipu’s GLM and MiniMax shipping frontier-class open models under MIT and Apache 2.0 licenses on a cadence Western labs haven’t matched.
The usage data reflects it: by May 2026, Chinese open-weight models accounted for roughly 61% of all tokens consumed on OpenRouter, with four of the five most-used models coming from Chinese labs — while Meta’s Llama, the prior open-weight leader, fell off the rankings entirely.
The U.S. counterexamples remain countable on one hand: OpenAI’s gpt-oss-120b and gpt-oss-20b, released under Apache 2.0 in August 2025 as the company’s first open weights since GPT-2; Google’s Gemma family, which is open-weight but ships under Google’s own more restrictive custom license rather than an OSI-approved one; and Thinking Machines’ Inkling.
Glimmer invites the most direct comparison to gpt-oss: both are Apache 2.0, both offer adjustable reasoning effort, and both target self-hosted deployment.
But the gpt-oss models are text-only, sparse mixture-of-experts designs built primarily for reasoning and tool use — gpt-oss-20b fits in about 16GB of memory while gpt-oss-120b targets a single 80GB data center GPU.
Glimmer stakes out different ground: a dense model with native vision input, trained end-to-end around the agent loop, shipping with its own quantized variants and speculative-decoding drafter tuned for 24GB consumer machines.
And if Zuckerberg follows through on opening Muse Spark 1.2’s weights, Meta would put an actual U.S. flagship frontier model into open circulation — something no American lab has done at that tier.
Safety remains part of the deployment architecture
Giving a local model access to tools creates a different security problem from deploying a local chatbot — and Meta’s own safety numbers show Glimmer is not uniformly stronger than its peers.
On CI Memories, a privacy benchmark where lower violation rates are better, Glimmer records 26.4 against Gemma’s 12.1 and Qwen’s 53.4. On Siren AgentDojo, a prompt-injection test, Glimmer shows a 28.4% attack-success rate versus 25.6% for Gemma and 40.3% for Qwen — while posting the highest utility score of the three at 94.2.
Meta says it evaluated Glimmer under its Advanced AI Scaling Framework and determined the model does not meet the framework’s definition of «Frontier AI» because it is generally less capable than Muse Spark. Its Preparedness Team assessed Glimmer at Moderate or lower risk across chemical/biological, cyber and loss-of-control categories — the latter two inferred from the fact that Glimmer is broadly weaker than Muse Spark 1.0, which received the same designations.
The company nevertheless recommends deploying Glimmer as part of a broader system with guardrails, including human-in-the-loop confirmation for irreversible actions. That caveat matters especially for local agents: keeping data on-device reduces exposure to cloud infrastructure, but local execution does not by itself solve prompt injection, excessive permissions or an agent taking an unintended action.
Apache 2.0 weights and a fast-growing runtime ecosystem
Meta is releasing full-precision BF16 weights, both 4-bit quantized variants, the DFlash drafter and the perception encoder — all under Apache 2.0. There is no Meta API price attached to the downloadable model, leaving total cost dependent on local hardware or whatever third-party hosting developers choose. One nuance worth noting for procurement teams: as with most «open source» model releases, it is the weights that are open — Meta has not released the training data or training code.
The broader implication is that Meta is treating the developer workstation as a credible deployment target for autonomous agents, rather than merely a place to experiment with smaller language models. Glimmer’s 30B size and 24GB target put that proposition within reach of high-end consumer hardware, while the Apache 2.0 license gives developers — and their legal departments — unusual freedom to modify and deploy it.
The next test is whether its benchmark advantages survive the messier conditions of real software repositories, enterprise tools and long-running agent sessions. If they do, the most consequential part of Glimmer may not be another set of benchmark scores — it may be that a class of agent previously expected to live behind a cloud API can increasingly live, and work, on the machine sitting under a developer’s desk.
Meta today released Muse Glimmer, a 30-billion-parameter open-weight model designed to run autonomous AI agents directly on consumer hardware — pushing agentic workloads that normally depend on cloud infrastructure onto high-end Macs and PCs.
Just as notable as what the model does is how it’s licensed. Glimmer arrives under the permissive, industry-standard Apache 2.0 open source license — the company’s first fully open release since it succeeded its open-weight Llama family in April with the proprietary Muse Spark.
In fact, Muse Glimmer launches today with a more permissive license than Llama ever carried. Llama’s bespoke community license drew years of criticism for restrictions like its 700-million-monthly-user cutoff; Apache 2.0 has no such strings, permitting unrestricted commercial use, modification and redistribution.
The weights are available on Hugging Face now. Wang said support is rolling out this week through Ollama, LM Studio, vLLM, SGLang, Together AI, Fireworks AI and OpenRouter, with optimized llama.cpp, MLX and ExecuTorch integrations landing in the coming days; Meta’s blog post also names Unsloth as a local-runtime partner and points to PyTorch’s TorchTitan for fine-tuning. The company says it is working with AMD, Arm, Dell, Intel and Nvidia to optimize performance across devices, and has published developer documentation covering custom agent scaffolds.
«Today we’re also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally,» Meta co-founder and CEO Mark Zuckerberg wrote in a post on X (under his longtime handle @finkd). «Soon we’ll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I’m proud of these releases.»
That promised Muse Spark 1.2 release would be an even bigger shift: it’s the frontier model behind Muse Code, the terminal coding agent Meta shipped just five days ago, and until today the entire Muse family was proprietary. Zuckerberg had teased at that launch that he’d «have more to share soon» on open source. Now we know what he meant.
For developers and enterprises, the practical stakes of local inference go beyond where computation happens. An agent working with files, screenshots, development environments and other sensitive context can execute those workflows without continuously sending that information to a remote inference service. Local deployment also removes network availability and per-token API charges from the inference loop — although organizations still bear hardware, electricity, deployment and management costs.
A 30B model built around the agent loop
Rather than positioning Glimmer primarily as a general chatbot, Meta trained it around the sequence of operations an autonomous agent performs: formulate a plan, call tools, interpret the results, continue working, and recover when something goes wrong.
«Just like much larger models, muse glimmer can operate as a fully capable agent via planning, tool calls, checking its own results, and failure recovery,» Alexandr Wang, Meta’s chief AI officer, wrote in a thread on X announcing the release, adding that the model «can run on 24GB of VRAM without losing agentic reliability.»
According to the model card on Hugging Face, Glimmer is a dense causal transformer with approximately 29.6 billion total parameters across 52 layers, including a dedicated ~1.8B-parameter ViT-G/14 perception encoder. It accepts interleaved text and images, produces text, supports more than 100 languages and has a stated context length of 131,072 tokens or more, with a knowledge cutoff of January 4, 2026.
That combination is intended to let an agent interpret screenshots, charts and documents while simultaneously reasoning about text and invoking external tools. Glimmer offers low, medium, high and xhigh reasoning settings — set via the system prompt — so applications can dial reasoning effort up or down per task, and Meta says it works across agentic scaffolds including OpenClaw and Hermes Agent.
The model is a distillation of Meta’s larger flagship: per the company’s technical blog post, Glimmer was pre-trained on Muse Spark’s outputs using logit distillation, mid-trained on longer-context, agent-heavy data with richer reasoning traces, then post-trained with supervised fine-tuning, on-policy distillation and reinforcement learning across general, reasoning, coding and agentic domains.
Meta demonstrated the result with a local Home Assistant workflow: in a demo video, Glimmer autonomously discovers a Home Assistant instance on the network via tool calls, queries device APIs, writes a responsive HTML/CSS/JavaScript dashboard from scratch and deploys a local server to verify its own work. That’s closer to the operational reality of enterprise agent deployments than a standalone question-answering benchmark — the model has to maintain a plan while interacting with external systems, then inspect whether its actions produced the expected result.
Compressing an agent into 24GB
The hardware story is central to the release.
At full precision, Meta says the 30B model requires more than 55GB of memory — beyond any single consumer GPU.
The company therefore developed approximately 4-bit quantized versions that shrink the language-model weights to under 20GB, leaving headroom for the pieces an operational agent also needs in memory: the KV cache, the perception encoder and a companion speculative-decoding model, all fitting within a 24GB or 32GB envelope.
In practical terms, that means the quantized builds run on consumer machines — though the upper end of them. The 24GB-targeted K-Quant-17GB configuration fits on a single high-end consumer graphics card, such as Nvidia’s RTX 3090 or RTX 4090 (both with 24GB of VRAM), while the 32GB-targeted K-Quant-Dynamic version lines up with the newer RTX 5090’s 32GB. On the Mac side, Apple Silicon’s unified memory plays the role of VRAM, so a MacBook Pro or Mac Studio with 32GB or more of memory can hold the full stack — Meta ran its own speed tests on M4 Max and M5 Max MacBook Pros. A typical 8GB or 16GB laptop, however, remains out of reach, and the full-precision BF16 release — which Meta pegs at 64GB — stays in the territory of data-center GPUs and top-spec Mac Studio configurations.
Meta reports average accuracy degradation of just 0.2% across 15 benchmarks for its K-Quant-Dynamic version targeting 32GB hardware, and 1% for the K-Quant-17GB configuration targeting 24GB hardware. Those figures are Meta’s own measurements, not independent evaluations.
Meta is also using DFlash speculative decoding to attack the other big problem with local agents: latency. Instead of generating every token sequentially, a smaller DFlash «drafter» model proposes blocks of 16 tokens that the primary model verifies in parallel, producing identical output faster.
Meta reports this raises average generation speed on an Nvidia RTX 5090 from 74.9 tokens per second to 233.4 — a 3.1x increase. An Apple M5 Max rises from 26.6 to 50.2 tokens per second (1.8x), and an M4 Max from 23.7 to 37.8 (1.5x). The tests used batch size one and greedy decoding, with Apple systems measured via ExecuTorch and the RTX 5090 via llama.cpp.
For agent applications, those multipliers matter more than they would for chat: a single user request can trigger many model turns, tool calls and verification steps, and latency accumulated at every stage can quickly make an otherwise capable agent impractical.
Glimmer enters an increasingly competitive local-model market
Meta is not entering an empty field. Developers already have capable open-weight models in this size class, most prominently Google’s Gemma 4 family and Alibaba’s Qwen3.6-27B — both of which position themselves around reasoning, multimodal understanding and agentic workloads. Meta’s own benchmark table compares directly against both.
Glimmer leads that three-way comparison on several agentic tests, including MCP Atlas at 75.5, DeepSearch QA at 74.6, τ³-Banking at 23.5, WildClawBench at 47.6 and GAIA2 at 43.3. It scores 51.2 on SWE-Bench Pro, versus 36.9 for Gemma4-31B and 50.2 for Qwen3.6-27B in Meta’s evaluation.
But Glimmer does not sweep the field. Qwen leads Meta’s own comparison on OSWorld-Verified (75.6 vs. Glimmer’s 65.9), TerminalBench 2.1 (60.7 vs. 51.7), SkillsBench, GDPval-AA (1141 vs. 953) and most of the multimodal benchmarks. On SWE-Bench Verified, Glimmer’s 76.0 lands just below Qwen’s 77.2. Gemma leads on GPQA Diamond and Humanity’s Last Exam.
Read honestly, the numbers make Glimmer more interesting as a specialized local-agent model than as evidence of a universal performance lead. For enterprise developers, the practical question is whether its combination of agent reliability, quantization quality, tool compatibility and decoding speed translates from benchmarks into sustained real-world workflows.
|
Model |
Developer / origin |
AA score |
Parameters / context |
Lowest tracked API price |
Access |
License |
Strongest use cases |
|
Moonshot AI; China |
60 |
2.8T total / 104B active; 1M |
$3.00 input / $15.00 output via Kimi, Fireworks or Modal (pricing) |
Custom Kimi K3 license. Large model-as-a-service operators above $20M in 12-month revenue need a separate agreement Large products may need to display “Kimi K3.” |
Frontier long-horizon coding Multimodal research and complex tool-driven agents |
||
|
Z.ai / Zhipu AI; China |
53 |
753B / 40B active; 1M |
$0.75 / $2.40 via DeepInfra FP4 (pricing) |
MIT |
Long-horizon coding and agents Million-token analysis with adjustable reasoning |
||
|
DeepSeek; China |
52 |
284B / 13B active; 1M |
$0.09 / $0.18 via DeepInfra (pricing) |
MIT |
• Extremely economical reasoning• Coding agents, terminal work and tool use |
||
|
MiniMax; China |
45 |
428B / 23B active; 1M |
$0.23 / $0.96 via CoreWeave (pricing) |
; |
MiniMax Community License. Commercial attribution required; companies above $20M yearly revenue need authorization. Includes prohibited-use conditions. |
• Native text, image and video work• Long-context coding and “cowork” agents |
|
|
Xiaomi; China |
43 |
1.02T / 42B active; 1M |
$0.35 / $0.70 via GMI (pricing) |
MIT |
Complex software engineering Agents spanning thousands of tool calls |
||
|
Thinking Machines Lab; U.S. |
42 |
975B / 41B active; 1M in weights |
$0.95 / $4.05 via DeepInfra FP8 (pricing) |
Apache 2.0 |
Customizable text, image and audio foundation Fine-tuned coding, RAG and tool-use systems |
||
|
NVIDIA; U.S. |
38 |
550B / 55B active; up to 1M in weights |
$0.37 / $1.08 via Blackbox AI (pricing) |
OpenMDW-1.1; permissive commercial and derivative-model rights |
Complex agents and long-context reasoning High-accuracy RAG, code, math and science |
||
|
Mistral AI; France |
30 |
128B dense; 256K |
$1.50 / $7.50 via Mistral (pricing) |
Modified MIT. Companies above $20M consolidated monthly revenue must obtain a commercial license or use Mistral’s service. |
Coding agents and function calling Multimodal instruction following |
||
|
Google DeepMind; U.S. |
30 |
30.7B dense; 256K |
Free on Google AI Studio’s limited tier; paid low $0.10 / $0.34 via CoreWeave (pricing) |
Apache 2.0 |
Compact multimodal reasoning and coding Manageable local or private-server deployments |
||
|
OpenAI; U.S. |
24 |
117B / 5.1B active; 131K |
$0.03 / $0.17 via CoreWeave (pricing) |
Numerous third-party APIs |
Apache 2.0 |
Reasoning structured output and tools Fine-tuning and single-80GB-GPU deployment |
|
|
Cohere; Canada |
23 |
218B / 25B active; 128K input |
Free on Cohere’s currently tracked endpoint (pricing) |
Apache 2.0 |
Enterprise RAG and grounded citations• Multilingual agents and document processing |
||
|
Meta; U.S. |
Not yet scored |
29.6B dense, including vision encoder; 131K+ |
No public metered hosted price located on launch day |
Apache 2.0 for full-precision weights, quantizations, drafter and perception encoder |
Always-on local agents on 24–32GB systems Tool use, recovery, coding and screen/document understanding |
Meta Glimmer adds to a still-small roster of genuinely open, frontier-class models from U.S. companies.
For the last two years, Chinese companies have set the pace in open source AI, with DeepSeek, Alibaba’s Qwen team, Moonshot AI’s Kimi, Zhipu’s GLM and MiniMax shipping frontier-class open models under MIT and Apache 2.0 licenses on a cadence Western labs haven’t matched.
The usage data reflects it: by May 2026, Chinese open-weight models accounted for roughly 61% of all tokens consumed on OpenRouter, with four of the five most-used models coming from Chinese labs — while Meta’s Llama, the prior open-weight leader, fell off the rankings entirely.
The U.S. counterexamples remain countable on one hand: OpenAI’s gpt-oss-120b and gpt-oss-20b, released under Apache 2.0 in August 2025 as the company’s first open weights since GPT-2; Google’s Gemma family, which is open-weight but ships under Google’s own more restrictive custom license rather than an OSI-approved one; and Thinking Machines’ Inkling.
Glimmer invites the most direct comparison to gpt-oss: both are Apache 2.0, both offer adjustable reasoning effort, and both target self-hosted deployment.
But the gpt-oss models are text-only, sparse mixture-of-experts designs built primarily for reasoning and tool use — gpt-oss-20b fits in about 16GB of memory while gpt-oss-120b targets a single 80GB data center GPU.
Glimmer stakes out different ground: a dense model with native vision input, trained end-to-end around the agent loop, shipping with its own quantized variants and speculative-decoding drafter tuned for 24GB consumer machines.
And if Zuckerberg follows through on opening Muse Spark 1.2’s weights, Meta would put an actual U.S. flagship frontier model into open circulation — something no American lab has done at that tier.
Safety remains part of the deployment architecture
Giving a local model access to tools creates a different security problem from deploying a local chatbot — and Meta’s own safety numbers show Glimmer is not uniformly stronger than its peers.
On CI Memories, a privacy benchmark where lower violation rates are better, Glimmer records 26.4 against Gemma’s 12.1 and Qwen’s 53.4. On Siren AgentDojo, a prompt-injection test, Glimmer shows a 28.4% attack-success rate versus 25.6% for Gemma and 40.3% for Qwen — while posting the highest utility score of the three at 94.2.
Meta says it evaluated Glimmer under its Advanced AI Scaling Framework and determined the model does not meet the framework’s definition of «Frontier AI» because it is generally less capable than Muse Spark. Its Preparedness Team assessed Glimmer at Moderate or lower risk across chemical/biological, cyber and loss-of-control categories — the latter two inferred from the fact that Glimmer is broadly weaker than Muse Spark 1.0, which received the same designations.
The company nevertheless recommends deploying Glimmer as part of a broader system with guardrails, including human-in-the-loop confirmation for irreversible actions. That caveat matters especially for local agents: keeping data on-device reduces exposure to cloud infrastructure, but local execution does not by itself solve prompt injection, excessive permissions or an agent taking an unintended action.
Apache 2.0 weights and a fast-growing runtime ecosystem
Meta is releasing full-precision BF16 weights, both 4-bit quantized variants, the DFlash drafter and the perception encoder — all under Apache 2.0. There is no Meta API price attached to the downloadable model, leaving total cost dependent on local hardware or whatever third-party hosting developers choose. One nuance worth noting for procurement teams: as with most «open source» model releases, it is the weights that are open — Meta has not released the training data or training code.
The broader implication is that Meta is treating the developer workstation as a credible deployment target for autonomous agents, rather than merely a place to experiment with smaller language models. Glimmer’s 30B size and 24GB target put that proposition within reach of high-end consumer hardware, while the Apache 2.0 license gives developers — and their legal departments — unusual freedom to modify and deploy it.
The next test is whether its benchmark advantages survive the messier conditions of real software repositories, enterprise tools and long-running agent sessions. If they do, the most consequential part of Glimmer may not be another set of benchmark scores — it may be that a class of agent previously expected to live behind a cloud API can increasingly live, and work, on the machine sitting under a developer’s desk.
Meta today released Muse Glimmer, a 30-billion-parameter open-weight model designed to run autonomous AI agents directly on consumer hardware — pushing agentic workloads that normally depend on cloud infrastructure onto high-end Macs and PCs.
Just as notable as what the model does is how it’s licensed. Glimmer arrives under the permissive, industry-standard Apache 2.0 open source license — the company’s first fully open release since it succeeded its open-weight Llama family in April with the proprietary Muse Spark.
In fact, Muse Glimmer launches today with a more permissive license than Llama ever carried. Llama’s bespoke community license drew years of criticism for restrictions like its 700-million-monthly-user cutoff; Apache 2.0 has no such strings, permitting unrestricted commercial use, modification and redistribution.
The weights are available on Hugging Face now. Wang said support is rolling out this week through Ollama, LM Studio, vLLM, SGLang, Together AI, Fireworks AI and OpenRouter, with optimized llama.cpp, MLX and ExecuTorch integrations landing in the coming days; Meta’s blog post also names Unsloth as a local-runtime partner and points to PyTorch’s TorchTitan for fine-tuning. The company says it is working with AMD, Arm, Dell, Intel and Nvidia to optimize performance across devices, and has published developer documentation covering custom agent scaffolds.
«Today we’re also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally,» Meta co-founder and CEO Mark Zuckerberg wrote in a post on X (under his longtime handle @finkd). «Soon we’ll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I’m proud of these releases.»
That promised Muse Spark 1.2 release would be an even bigger shift: it’s the frontier model behind Muse Code, the terminal coding agent Meta shipped just five days ago, and until today the entire Muse family was proprietary. Zuckerberg had teased at that launch that he’d «have more to share soon» on open source. Now we know what he meant.
For developers and enterprises, the practical stakes of local inference go beyond where computation happens. An agent working with files, screenshots, development environments and other sensitive context can execute those workflows without continuously sending that information to a remote inference service. Local deployment also removes network availability and per-token API charges from the inference loop — although organizations still bear hardware, electricity, deployment and management costs.
A 30B model built around the agent loop
Rather than positioning Glimmer primarily as a general chatbot, Meta trained it around the sequence of operations an autonomous agent performs: formulate a plan, call tools, interpret the results, continue working, and recover when something goes wrong.
«Just like much larger models, muse glimmer can operate as a fully capable agent via planning, tool calls, checking its own results, and failure recovery,» Alexandr Wang, Meta’s chief AI officer, wrote in a thread on X announcing the release, adding that the model «can run on 24GB of VRAM without losing agentic reliability.»
According to the model card on Hugging Face, Glimmer is a dense causal transformer with approximately 29.6 billion total parameters across 52 layers, including a dedicated ~1.8B-parameter ViT-G/14 perception encoder. It accepts interleaved text and images, produces text, supports more than 100 languages and has a stated context length of 131,072 tokens or more, with a knowledge cutoff of January 4, 2026.
That combination is intended to let an agent interpret screenshots, charts and documents while simultaneously reasoning about text and invoking external tools. Glimmer offers low, medium, high and xhigh reasoning settings — set via the system prompt — so applications can dial reasoning effort up or down per task, and Meta says it works across agentic scaffolds including OpenClaw and Hermes Agent.
The model is a distillation of Meta’s larger flagship: per the company’s technical blog post, Glimmer was pre-trained on Muse Spark’s outputs using logit distillation, mid-trained on longer-context, agent-heavy data with richer reasoning traces, then post-trained with supervised fine-tuning, on-policy distillation and reinforcement learning across general, reasoning, coding and agentic domains.
Meta demonstrated the result with a local Home Assistant workflow: in a demo video, Glimmer autonomously discovers a Home Assistant instance on the network via tool calls, queries device APIs, writes a responsive HTML/CSS/JavaScript dashboard from scratch and deploys a local server to verify its own work. That’s closer to the operational reality of enterprise agent deployments than a standalone question-answering benchmark — the model has to maintain a plan while interacting with external systems, then inspect whether its actions produced the expected result.
Compressing an agent into 24GB
The hardware story is central to the release.
At full precision, Meta says the 30B model requires more than 55GB of memory — beyond any single consumer GPU.
The company therefore developed approximately 4-bit quantized versions that shrink the language-model weights to under 20GB, leaving headroom for the pieces an operational agent also needs in memory: the KV cache, the perception encoder and a companion speculative-decoding model, all fitting within a 24GB or 32GB envelope.
In practical terms, that means the quantized builds run on consumer machines — though the upper end of them. The 24GB-targeted K-Quant-17GB configuration fits on a single high-end consumer graphics card, such as Nvidia’s RTX 3090 or RTX 4090 (both with 24GB of VRAM), while the 32GB-targeted K-Quant-Dynamic version lines up with the newer RTX 5090’s 32GB. On the Mac side, Apple Silicon’s unified memory plays the role of VRAM, so a MacBook Pro or Mac Studio with 32GB or more of memory can hold the full stack — Meta ran its own speed tests on M4 Max and M5 Max MacBook Pros. A typical 8GB or 16GB laptop, however, remains out of reach, and the full-precision BF16 release — which Meta pegs at 64GB — stays in the territory of data-center GPUs and top-spec Mac Studio configurations.
Meta reports average accuracy degradation of just 0.2% across 15 benchmarks for its K-Quant-Dynamic version targeting 32GB hardware, and 1% for the K-Quant-17GB configuration targeting 24GB hardware. Those figures are Meta’s own measurements, not independent evaluations.
Meta is also using DFlash speculative decoding to attack the other big problem with local agents: latency. Instead of generating every token sequentially, a smaller DFlash «drafter» model proposes blocks of 16 tokens that the primary model verifies in parallel, producing identical output faster.
Meta reports this raises average generation speed on an Nvidia RTX 5090 from 74.9 tokens per second to 233.4 — a 3.1x increase. An Apple M5 Max rises from 26.6 to 50.2 tokens per second (1.8x), and an M4 Max from 23.7 to 37.8 (1.5x). The tests used batch size one and greedy decoding, with Apple systems measured via ExecuTorch and the RTX 5090 via llama.cpp.
For agent applications, those multipliers matter more than they would for chat: a single user request can trigger many model turns, tool calls and verification steps, and latency accumulated at every stage can quickly make an otherwise capable agent impractical.
Glimmer enters an increasingly competitive local-model market
Meta is not entering an empty field. Developers already have capable open-weight models in this size class, most prominently Google’s Gemma 4 family and Alibaba’s Qwen3.6-27B — both of which position themselves around reasoning, multimodal understanding and agentic workloads. Meta’s own benchmark table compares directly against both.
Glimmer leads that three-way comparison on several agentic tests, including MCP Atlas at 75.5, DeepSearch QA at 74.6, τ³-Banking at 23.5, WildClawBench at 47.6 and GAIA2 at 43.3. It scores 51.2 on SWE-Bench Pro, versus 36.9 for Gemma4-31B and 50.2 for Qwen3.6-27B in Meta’s evaluation.
But Glimmer does not sweep the field. Qwen leads Meta’s own comparison on OSWorld-Verified (75.6 vs. Glimmer’s 65.9), TerminalBench 2.1 (60.7 vs. 51.7), SkillsBench, GDPval-AA (1141 vs. 953) and most of the multimodal benchmarks. On SWE-Bench Verified, Glimmer’s 76.0 lands just below Qwen’s 77.2. Gemma leads on GPQA Diamond and Humanity’s Last Exam.
Read honestly, the numbers make Glimmer more interesting as a specialized local-agent model than as evidence of a universal performance lead. For enterprise developers, the practical question is whether its combination of agent reliability, quantization quality, tool compatibility and decoding speed translates from benchmarks into sustained real-world workflows.
|
Model |
Developer / origin |
AA score |
Parameters / context |
Lowest tracked API price |
Access |
License |
Strongest use cases |
|
Moonshot AI; China |
60 |
2.8T total / 104B active; 1M |
$3.00 input / $15.00 output via Kimi, Fireworks or Modal (pricing) |
Custom Kimi K3 license. Large model-as-a-service operators above $20M in 12-month revenue need a separate agreement Large products may need to display “Kimi K3.” |
Frontier long-horizon coding Multimodal research and complex tool-driven agents |
||
|
Z.ai / Zhipu AI; China |
53 |
753B / 40B active; 1M |
$0.75 / $2.40 via DeepInfra FP4 (pricing) |
MIT |
Long-horizon coding and agents Million-token analysis with adjustable reasoning |
||
|
DeepSeek; China |
52 |
284B / 13B active; 1M |
$0.09 / $0.18 via DeepInfra (pricing) |
MIT |
• Extremely economical reasoning• Coding agents, terminal work and tool use |
||
|
MiniMax; China |
45 |
428B / 23B active; 1M |
$0.23 / $0.96 via CoreWeave (pricing) |
; |
MiniMax Community License. Commercial attribution required; companies above $20M yearly revenue need authorization. Includes prohibited-use conditions. |
• Native text, image and video work• Long-context coding and “cowork” agents |
|
|
Xiaomi; China |
43 |
1.02T / 42B active; 1M |
$0.35 / $0.70 via GMI (pricing) |
MIT |
Complex software engineering Agents spanning thousands of tool calls |
||
|
Thinking Machines Lab; U.S. |
42 |
975B / 41B active; 1M in weights |
$0.95 / $4.05 via DeepInfra FP8 (pricing) |
Apache 2.0 |
Customizable text, image and audio foundation Fine-tuned coding, RAG and tool-use systems |
||
|
NVIDIA; U.S. |
38 |
550B / 55B active; up to 1M in weights |
$0.37 / $1.08 via Blackbox AI (pricing) |
OpenMDW-1.1; permissive commercial and derivative-model rights |
Complex agents and long-context reasoning High-accuracy RAG, code, math and science |
||
|
Mistral AI; France |
30 |
128B dense; 256K |
$1.50 / $7.50 via Mistral (pricing) |
Modified MIT. Companies above $20M consolidated monthly revenue must obtain a commercial license or use Mistral’s service. |
Coding agents and function calling Multimodal instruction following |
||
|
Google DeepMind; U.S. |
30 |
30.7B dense; 256K |
Free on Google AI Studio’s limited tier; paid low $0.10 / $0.34 via CoreWeave (pricing) |
Apache 2.0 |
Compact multimodal reasoning and coding Manageable local or private-server deployments |
||
|
OpenAI; U.S. |
24 |
117B / 5.1B active; 131K |
$0.03 / $0.17 via CoreWeave (pricing) |
Numerous third-party APIs |
Apache 2.0 |
Reasoning structured output and tools Fine-tuning and single-80GB-GPU deployment |
|
|
Cohere; Canada |
23 |
218B / 25B active; 128K input |
Free on Cohere’s currently tracked endpoint (pricing) |
Apache 2.0 |
Enterprise RAG and grounded citations• Multilingual agents and document processing |
||
|
Meta; U.S. |
Not yet scored |
29.6B dense, including vision encoder; 131K+ |
No public metered hosted price located on launch day |
Apache 2.0 for full-precision weights, quantizations, drafter and perception encoder |
Always-on local agents on 24–32GB systems Tool use, recovery, coding and screen/document understanding |
Meta Glimmer adds to a still-small roster of genuinely open, frontier-class models from U.S. companies.
For the last two years, Chinese companies have set the pace in open source AI, with DeepSeek, Alibaba’s Qwen team, Moonshot AI’s Kimi, Zhipu’s GLM and MiniMax shipping frontier-class open models under MIT and Apache 2.0 licenses on a cadence Western labs haven’t matched.
The usage data reflects it: by May 2026, Chinese open-weight models accounted for roughly 61% of all tokens consumed on OpenRouter, with four of the five most-used models coming from Chinese labs — while Meta’s Llama, the prior open-weight leader, fell off the rankings entirely.
The U.S. counterexamples remain countable on one hand: OpenAI’s gpt-oss-120b and gpt-oss-20b, released under Apache 2.0 in August 2025 as the company’s first open weights since GPT-2; Google’s Gemma family, which is open-weight but ships under Google’s own more restrictive custom license rather than an OSI-approved one; and Thinking Machines’ Inkling.
Glimmer invites the most direct comparison to gpt-oss: both are Apache 2.0, both offer adjustable reasoning effort, and both target self-hosted deployment.
But the gpt-oss models are text-only, sparse mixture-of-experts designs built primarily for reasoning and tool use — gpt-oss-20b fits in about 16GB of memory while gpt-oss-120b targets a single 80GB data center GPU.
Glimmer stakes out different ground: a dense model with native vision input, trained end-to-end around the agent loop, shipping with its own quantized variants and speculative-decoding drafter tuned for 24GB consumer machines.
And if Zuckerberg follows through on opening Muse Spark 1.2’s weights, Meta would put an actual U.S. flagship frontier model into open circulation — something no American lab has done at that tier.
Safety remains part of the deployment architecture
Giving a local model access to tools creates a different security problem from deploying a local chatbot — and Meta’s own safety numbers show Glimmer is not uniformly stronger than its peers.
On CI Memories, a privacy benchmark where lower violation rates are better, Glimmer records 26.4 against Gemma’s 12.1 and Qwen’s 53.4. On Siren AgentDojo, a prompt-injection test, Glimmer shows a 28.4% attack-success rate versus 25.6% for Gemma and 40.3% for Qwen — while posting the highest utility score of the three at 94.2.
Meta says it evaluated Glimmer under its Advanced AI Scaling Framework and determined the model does not meet the framework’s definition of «Frontier AI» because it is generally less capable than Muse Spark. Its Preparedness Team assessed Glimmer at Moderate or lower risk across chemical/biological, cyber and loss-of-control categories — the latter two inferred from the fact that Glimmer is broadly weaker than Muse Spark 1.0, which received the same designations.
The company nevertheless recommends deploying Glimmer as part of a broader system with guardrails, including human-in-the-loop confirmation for irreversible actions. That caveat matters especially for local agents: keeping data on-device reduces exposure to cloud infrastructure, but local execution does not by itself solve prompt injection, excessive permissions or an agent taking an unintended action.
Apache 2.0 weights and a fast-growing runtime ecosystem
Meta is releasing full-precision BF16 weights, both 4-bit quantized variants, the DFlash drafter and the perception encoder — all under Apache 2.0. There is no Meta API price attached to the downloadable model, leaving total cost dependent on local hardware or whatever third-party hosting developers choose. One nuance worth noting for procurement teams: as with most «open source» model releases, it is the weights that are open — Meta has not released the training data or training code.
The broader implication is that Meta is treating the developer workstation as a credible deployment target for autonomous agents, rather than merely a place to experiment with smaller language models. Glimmer’s 30B size and 24GB target put that proposition within reach of high-end consumer hardware, while the Apache 2.0 license gives developers — and their legal departments — unusual freedom to modify and deploy it.
The next test is whether its benchmark advantages survive the messier conditions of real software repositories, enterprise tools and long-running agent sessions. If they do, the most consequential part of Glimmer may not be another set of benchmark scores — it may be that a class of agent previously expected to live behind a cloud API can increasingly live, and work, on the machine sitting under a developer’s desk.
Suscríbete y recibe las historias más importantes del día.
Al suscribirte aceptas nuestros términos y condiciones y política de privacidad.
Meta today released Muse Glimmer, a 30-billion-parameter open-weight model designed to run autonomous AI agents directly on consumer hardware — pushing agentic workloads that normally depend on cloud infrastructure onto high-end Macs and PCs.
Just as notable as what the model does is how it’s licensed. Glimmer arrives under the permissive, industry-standard Apache 2.0 open source license — the company’s first fully open release since it succeeded its open-weight Llama family in April with the proprietary Muse Spark.
In fact, Muse Glimmer launches today with a more permissive license than Llama ever carried. Llama’s bespoke community license drew years of criticism for restrictions like its 700-million-monthly-user cutoff; Apache 2.0 has no such strings, permitting unrestricted commercial use, modification and redistribution.
The weights are available on Hugging Face now. Wang said support is rolling out this week through Ollama, LM Studio, vLLM, SGLang, Together AI, Fireworks AI and OpenRouter, with optimized llama.cpp, MLX and ExecuTorch integrations landing in the coming days; Meta’s blog post also names Unsloth as a local-runtime partner and points to PyTorch’s TorchTitan for fine-tuning. The company says it is working with AMD, Arm, Dell, Intel and Nvidia to optimize performance across devices, and has published developer documentation covering custom agent scaffolds.
«Today we’re also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally,» Meta co-founder and CEO Mark Zuckerberg wrote in a post on X (under his longtime handle @finkd). «Soon we’ll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I’m proud of these releases.»
That promised Muse Spark 1.2 release would be an even bigger shift: it’s the frontier model behind Muse Code, the terminal coding agent Meta shipped just five days ago, and until today the entire Muse family was proprietary. Zuckerberg had teased at that launch that he’d «have more to share soon» on open source. Now we know what he meant.
For developers and enterprises, the practical stakes of local inference go beyond where computation happens. An agent working with files, screenshots, development environments and other sensitive context can execute those workflows without continuously sending that information to a remote inference service. Local deployment also removes network availability and per-token API charges from the inference loop — although organizations still bear hardware, electricity, deployment and management costs.
A 30B model built around the agent loop
Rather than positioning Glimmer primarily as a general chatbot, Meta trained it around the sequence of operations an autonomous agent performs: formulate a plan, call tools, interpret the results, continue working, and recover when something goes wrong.
«Just like much larger models, muse glimmer can operate as a fully capable agent via planning, tool calls, checking its own results, and failure recovery,» Alexandr Wang, Meta’s chief AI officer, wrote in a thread on X announcing the release, adding that the model «can run on 24GB of VRAM without losing agentic reliability.»
According to the model card on Hugging Face, Glimmer is a dense causal transformer with approximately 29.6 billion total parameters across 52 layers, including a dedicated ~1.8B-parameter ViT-G/14 perception encoder. It accepts interleaved text and images, produces text, supports more than 100 languages and has a stated context length of 131,072 tokens or more, with a knowledge cutoff of January 4, 2026.
That combination is intended to let an agent interpret screenshots, charts and documents while simultaneously reasoning about text and invoking external tools. Glimmer offers low, medium, high and xhigh reasoning settings — set via the system prompt — so applications can dial reasoning effort up or down per task, and Meta says it works across agentic scaffolds including OpenClaw and Hermes Agent.
The model is a distillation of Meta’s larger flagship: per the company’s technical blog post, Glimmer was pre-trained on Muse Spark’s outputs using logit distillation, mid-trained on longer-context, agent-heavy data with richer reasoning traces, then post-trained with supervised fine-tuning, on-policy distillation and reinforcement learning across general, reasoning, coding and agentic domains.
Meta demonstrated the result with a local Home Assistant workflow: in a demo video, Glimmer autonomously discovers a Home Assistant instance on the network via tool calls, queries device APIs, writes a responsive HTML/CSS/JavaScript dashboard from scratch and deploys a local server to verify its own work. That’s closer to the operational reality of enterprise agent deployments than a standalone question-answering benchmark — the model has to maintain a plan while interacting with external systems, then inspect whether its actions produced the expected result.
Compressing an agent into 24GB
The hardware story is central to the release.
At full precision, Meta says the 30B model requires more than 55GB of memory — beyond any single consumer GPU.
The company therefore developed approximately 4-bit quantized versions that shrink the language-model weights to under 20GB, leaving headroom for the pieces an operational agent also needs in memory: the KV cache, the perception encoder and a companion speculative-decoding model, all fitting within a 24GB or 32GB envelope.
In practical terms, that means the quantized builds run on consumer machines — though the upper end of them. The 24GB-targeted K-Quant-17GB configuration fits on a single high-end consumer graphics card, such as Nvidia’s RTX 3090 or RTX 4090 (both with 24GB of VRAM), while the 32GB-targeted K-Quant-Dynamic version lines up with the newer RTX 5090’s 32GB. On the Mac side, Apple Silicon’s unified memory plays the role of VRAM, so a MacBook Pro or Mac Studio with 32GB or more of memory can hold the full stack — Meta ran its own speed tests on M4 Max and M5 Max MacBook Pros. A typical 8GB or 16GB laptop, however, remains out of reach, and the full-precision BF16 release — which Meta pegs at 64GB — stays in the territory of data-center GPUs and top-spec Mac Studio configurations.
Meta reports average accuracy degradation of just 0.2% across 15 benchmarks for its K-Quant-Dynamic version targeting 32GB hardware, and 1% for the K-Quant-17GB configuration targeting 24GB hardware. Those figures are Meta’s own measurements, not independent evaluations.
Meta is also using DFlash speculative decoding to attack the other big problem with local agents: latency. Instead of generating every token sequentially, a smaller DFlash «drafter» model proposes blocks of 16 tokens that the primary model verifies in parallel, producing identical output faster.
Meta reports this raises average generation speed on an Nvidia RTX 5090 from 74.9 tokens per second to 233.4 — a 3.1x increase. An Apple M5 Max rises from 26.6 to 50.2 tokens per second (1.8x), and an M4 Max from 23.7 to 37.8 (1.5x). The tests used batch size one and greedy decoding, with Apple systems measured via ExecuTorch and the RTX 5090 via llama.cpp.
For agent applications, those multipliers matter more than they would for chat: a single user request can trigger many model turns, tool calls and verification steps, and latency accumulated at every stage can quickly make an otherwise capable agent impractical.
Glimmer enters an increasingly competitive local-model market
Meta is not entering an empty field. Developers already have capable open-weight models in this size class, most prominently Google’s Gemma 4 family and Alibaba’s Qwen3.6-27B — both of which position themselves around reasoning, multimodal understanding and agentic workloads. Meta’s own benchmark table compares directly against both.
Glimmer leads that three-way comparison on several agentic tests, including MCP Atlas at 75.5, DeepSearch QA at 74.6, τ³-Banking at 23.5, WildClawBench at 47.6 and GAIA2 at 43.3. It scores 51.2 on SWE-Bench Pro, versus 36.9 for Gemma4-31B and 50.2 for Qwen3.6-27B in Meta’s evaluation.
But Glimmer does not sweep the field. Qwen leads Meta’s own comparison on OSWorld-Verified (75.6 vs. Glimmer’s 65.9), TerminalBench 2.1 (60.7 vs. 51.7), SkillsBench, GDPval-AA (1141 vs. 953) and most of the multimodal benchmarks. On SWE-Bench Verified, Glimmer’s 76.0 lands just below Qwen’s 77.2. Gemma leads on GPQA Diamond and Humanity’s Last Exam.
Read honestly, the numbers make Glimmer more interesting as a specialized local-agent model than as evidence of a universal performance lead. For enterprise developers, the practical question is whether its combination of agent reliability, quantization quality, tool compatibility and decoding speed translates from benchmarks into sustained real-world workflows.
|
Model |
Developer / origin |
AA score |
Parameters / context |
Lowest tracked API price |
Access |
License |
Strongest use cases |
|
Moonshot AI; China |
60 |
2.8T total / 104B active; 1M |
$3.00 input / $15.00 output via Kimi, Fireworks or Modal (pricing) |
Custom Kimi K3 license. Large model-as-a-service operators above $20M in 12-month revenue need a separate agreement Large products may need to display “Kimi K3.” |
Frontier long-horizon coding Multimodal research and complex tool-driven agents |
||
|
Z.ai / Zhipu AI; China |
53 |
753B / 40B active; 1M |
$0.75 / $2.40 via DeepInfra FP4 (pricing) |
MIT |
Long-horizon coding and agents Million-token analysis with adjustable reasoning |
||
|
DeepSeek; China |
52 |
284B / 13B active; 1M |
$0.09 / $0.18 via DeepInfra (pricing) |
MIT |
• Extremely economical reasoning• Coding agents, terminal work and tool use |
||
|
MiniMax; China |
45 |
428B / 23B active; 1M |
$0.23 / $0.96 via CoreWeave (pricing) |
; |
MiniMax Community License. Commercial attribution required; companies above $20M yearly revenue need authorization. Includes prohibited-use conditions. |
• Native text, image and video work• Long-context coding and “cowork” agents |
|
|
Xiaomi; China |
43 |
1.02T / 42B active; 1M |
$0.35 / $0.70 via GMI (pricing) |
MIT |
Complex software engineering Agents spanning thousands of tool calls |
||
|
Thinking Machines Lab; U.S. |
42 |
975B / 41B active; 1M in weights |
$0.95 / $4.05 via DeepInfra FP8 (pricing) |
Apache 2.0 |
Customizable text, image and audio foundation Fine-tuned coding, RAG and tool-use systems |
||
|
NVIDIA; U.S. |
38 |
550B / 55B active; up to 1M in weights |
$0.37 / $1.08 via Blackbox AI (pricing) |
OpenMDW-1.1; permissive commercial and derivative-model rights |
Complex agents and long-context reasoning High-accuracy RAG, code, math and science |
||
|
Mistral AI; France |
30 |
128B dense; 256K |
$1.50 / $7.50 via Mistral (pricing) |
Modified MIT. Companies above $20M consolidated monthly revenue must obtain a commercial license or use Mistral’s service. |
Coding agents and function calling Multimodal instruction following |
||
|
Google DeepMind; U.S. |
30 |
30.7B dense; 256K |
Free on Google AI Studio’s limited tier; paid low $0.10 / $0.34 via CoreWeave (pricing) |
Apache 2.0 |
Compact multimodal reasoning and coding Manageable local or private-server deployments |
||
|
OpenAI; U.S. |
24 |
117B / 5.1B active; 131K |
$0.03 / $0.17 via CoreWeave (pricing) |
Numerous third-party APIs |
Apache 2.0 |
Reasoning structured output and tools Fine-tuning and single-80GB-GPU deployment |
|
|
Cohere; Canada |
23 |
218B / 25B active; 128K input |
Free on Cohere’s currently tracked endpoint (pricing) |
Apache 2.0 |
Enterprise RAG and grounded citations• Multilingual agents and document processing |
||
|
Meta; U.S. |
Not yet scored |
29.6B dense, including vision encoder; 131K+ |
No public metered hosted price located on launch day |
Apache 2.0 for full-precision weights, quantizations, drafter and perception encoder |
Always-on local agents on 24–32GB systems Tool use, recovery, coding and screen/document understanding |
Meta Glimmer adds to a still-small roster of genuinely open, frontier-class models from U.S. companies.
For the last two years, Chinese companies have set the pace in open source AI, with DeepSeek, Alibaba’s Qwen team, Moonshot AI’s Kimi, Zhipu’s GLM and MiniMax shipping frontier-class open models under MIT and Apache 2.0 licenses on a cadence Western labs haven’t matched.
The usage data reflects it: by May 2026, Chinese open-weight models accounted for roughly 61% of all tokens consumed on OpenRouter, with four of the five most-used models coming from Chinese labs — while Meta’s Llama, the prior open-weight leader, fell off the rankings entirely.
The U.S. counterexamples remain countable on one hand: OpenAI’s gpt-oss-120b and gpt-oss-20b, released under Apache 2.0 in August 2025 as the company’s first open weights since GPT-2; Google’s Gemma family, which is open-weight but ships under Google’s own more restrictive custom license rather than an OSI-approved one; and Thinking Machines’ Inkling.
Glimmer invites the most direct comparison to gpt-oss: both are Apache 2.0, both offer adjustable reasoning effort, and both target self-hosted deployment.
But the gpt-oss models are text-only, sparse mixture-of-experts designs built primarily for reasoning and tool use — gpt-oss-20b fits in about 16GB of memory while gpt-oss-120b targets a single 80GB data center GPU.
Glimmer stakes out different ground: a dense model with native vision input, trained end-to-end around the agent loop, shipping with its own quantized variants and speculative-decoding drafter tuned for 24GB consumer machines.
And if Zuckerberg follows through on opening Muse Spark 1.2’s weights, Meta would put an actual U.S. flagship frontier model into open circulation — something no American lab has done at that tier.
Safety remains part of the deployment architecture
Giving a local model access to tools creates a different security problem from deploying a local chatbot — and Meta’s own safety numbers show Glimmer is not uniformly stronger than its peers.
On CI Memories, a privacy benchmark where lower violation rates are better, Glimmer records 26.4 against Gemma’s 12.1 and Qwen’s 53.4. On Siren AgentDojo, a prompt-injection test, Glimmer shows a 28.4% attack-success rate versus 25.6% for Gemma and 40.3% for Qwen — while posting the highest utility score of the three at 94.2.
Meta says it evaluated Glimmer under its Advanced AI Scaling Framework and determined the model does not meet the framework’s definition of «Frontier AI» because it is generally less capable than Muse Spark. Its Preparedness Team assessed Glimmer at Moderate or lower risk across chemical/biological, cyber and loss-of-control categories — the latter two inferred from the fact that Glimmer is broadly weaker than Muse Spark 1.0, which received the same designations.
The company nevertheless recommends deploying Glimmer as part of a broader system with guardrails, including human-in-the-loop confirmation for irreversible actions. That caveat matters especially for local agents: keeping data on-device reduces exposure to cloud infrastructure, but local execution does not by itself solve prompt injection, excessive permissions or an agent taking an unintended action.
Apache 2.0 weights and a fast-growing runtime ecosystem
Meta is releasing full-precision BF16 weights, both 4-bit quantized variants, the DFlash drafter and the perception encoder — all under Apache 2.0. There is no Meta API price attached to the downloadable model, leaving total cost dependent on local hardware or whatever third-party hosting developers choose. One nuance worth noting for procurement teams: as with most «open source» model releases, it is the weights that are open — Meta has not released the training data or training code.
The broader implication is that Meta is treating the developer workstation as a credible deployment target for autonomous agents, rather than merely a place to experiment with smaller language models. Glimmer’s 30B size and 24GB target put that proposition within reach of high-end consumer hardware, while the Apache 2.0 license gives developers — and their legal departments — unusual freedom to modify and deploy it.
The next test is whether its benchmark advantages survive the messier conditions of real software repositories, enterprise tools and long-running agent sessions. If they do, the most consequential part of Glimmer may not be another set of benchmark scores — it may be that a class of agent previously expected to live behind a cloud API can increasingly live, and work, on the machine sitting under a developer’s desk.
Meta today released Muse Glimmer, a 30-billion-parameter open-weight model designed to run autonomous AI agents directly on consumer hardware — pushing agentic workloads that normally depend on cloud infrastructure onto high-end Macs and PCs.
Just as notable as what the model does is how it’s licensed. Glimmer arrives under the permissive, industry-standard Apache 2.0 open source license — the company’s first fully open release since it succeeded its open-weight Llama family in April with the proprietary Muse Spark.
In fact, Muse Glimmer launches today with a more permissive license than Llama ever carried. Llama’s bespoke community license drew years of criticism for restrictions like its 700-million-monthly-user cutoff; Apache 2.0 has no such strings, permitting unrestricted commercial use, modification and redistribution.
The weights are available on Hugging Face now. Wang said support is rolling out this week through Ollama, LM Studio, vLLM, SGLang, Together AI, Fireworks AI and OpenRouter, with optimized llama.cpp, MLX and ExecuTorch integrations landing in the coming days; Meta’s blog post also names Unsloth as a local-runtime partner and points to PyTorch’s TorchTitan for fine-tuning. The company says it is working with AMD, Arm, Dell, Intel and Nvidia to optimize performance across devices, and has published developer documentation covering custom agent scaffolds.
«Today we’re also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally,» Meta co-founder and CEO Mark Zuckerberg wrote in a post on X (under his longtime handle @finkd). «Soon we’ll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I’m proud of these releases.»
That promised Muse Spark 1.2 release would be an even bigger shift: it’s the frontier model behind Muse Code, the terminal coding agent Meta shipped just five days ago, and until today the entire Muse family was proprietary. Zuckerberg had teased at that launch that he’d «have more to share soon» on open source. Now we know what he meant.
For developers and enterprises, the practical stakes of local inference go beyond where computation happens. An agent working with files, screenshots, development environments and other sensitive context can execute those workflows without continuously sending that information to a remote inference service. Local deployment also removes network availability and per-token API charges from the inference loop — although organizations still bear hardware, electricity, deployment and management costs.
A 30B model built around the agent loop
Rather than positioning Glimmer primarily as a general chatbot, Meta trained it around the sequence of operations an autonomous agent performs: formulate a plan, call tools, interpret the results, continue working, and recover when something goes wrong.
«Just like much larger models, muse glimmer can operate as a fully capable agent via planning, tool calls, checking its own results, and failure recovery,» Alexandr Wang, Meta’s chief AI officer, wrote in a thread on X announcing the release, adding that the model «can run on 24GB of VRAM without losing agentic reliability.»
According to the model card on Hugging Face, Glimmer is a dense causal transformer with approximately 29.6 billion total parameters across 52 layers, including a dedicated ~1.8B-parameter ViT-G/14 perception encoder. It accepts interleaved text and images, produces text, supports more than 100 languages and has a stated context length of 131,072 tokens or more, with a knowledge cutoff of January 4, 2026.
That combination is intended to let an agent interpret screenshots, charts and documents while simultaneously reasoning about text and invoking external tools. Glimmer offers low, medium, high and xhigh reasoning settings — set via the system prompt — so applications can dial reasoning effort up or down per task, and Meta says it works across agentic scaffolds including OpenClaw and Hermes Agent.
The model is a distillation of Meta’s larger flagship: per the company’s technical blog post, Glimmer was pre-trained on Muse Spark’s outputs using logit distillation, mid-trained on longer-context, agent-heavy data with richer reasoning traces, then post-trained with supervised fine-tuning, on-policy distillation and reinforcement learning across general, reasoning, coding and agentic domains.
Meta demonstrated the result with a local Home Assistant workflow: in a demo video, Glimmer autonomously discovers a Home Assistant instance on the network via tool calls, queries device APIs, writes a responsive HTML/CSS/JavaScript dashboard from scratch and deploys a local server to verify its own work. That’s closer to the operational reality of enterprise agent deployments than a standalone question-answering benchmark — the model has to maintain a plan while interacting with external systems, then inspect whether its actions produced the expected result.
Compressing an agent into 24GB
The hardware story is central to the release.
At full precision, Meta says the 30B model requires more than 55GB of memory — beyond any single consumer GPU.
The company therefore developed approximately 4-bit quantized versions that shrink the language-model weights to under 20GB, leaving headroom for the pieces an operational agent also needs in memory: the KV cache, the perception encoder and a companion speculative-decoding model, all fitting within a 24GB or 32GB envelope.
In practical terms, that means the quantized builds run on consumer machines — though the upper end of them. The 24GB-targeted K-Quant-17GB configuration fits on a single high-end consumer graphics card, such as Nvidia’s RTX 3090 or RTX 4090 (both with 24GB of VRAM), while the 32GB-targeted K-Quant-Dynamic version lines up with the newer RTX 5090’s 32GB. On the Mac side, Apple Silicon’s unified memory plays the role of VRAM, so a MacBook Pro or Mac Studio with 32GB or more of memory can hold the full stack — Meta ran its own speed tests on M4 Max and M5 Max MacBook Pros. A typical 8GB or 16GB laptop, however, remains out of reach, and the full-precision BF16 release — which Meta pegs at 64GB — stays in the territory of data-center GPUs and top-spec Mac Studio configurations.
Meta reports average accuracy degradation of just 0.2% across 15 benchmarks for its K-Quant-Dynamic version targeting 32GB hardware, and 1% for the K-Quant-17GB configuration targeting 24GB hardware. Those figures are Meta’s own measurements, not independent evaluations.
Meta is also using DFlash speculative decoding to attack the other big problem with local agents: latency. Instead of generating every token sequentially, a smaller DFlash «drafter» model proposes blocks of 16 tokens that the primary model verifies in parallel, producing identical output faster.
Meta reports this raises average generation speed on an Nvidia RTX 5090 from 74.9 tokens per second to 233.4 — a 3.1x increase. An Apple M5 Max rises from 26.6 to 50.2 tokens per second (1.8x), and an M4 Max from 23.7 to 37.8 (1.5x). The tests used batch size one and greedy decoding, with Apple systems measured via ExecuTorch and the RTX 5090 via llama.cpp.
For agent applications, those multipliers matter more than they would for chat: a single user request can trigger many model turns, tool calls and verification steps, and latency accumulated at every stage can quickly make an otherwise capable agent impractical.
Glimmer enters an increasingly competitive local-model market
Meta is not entering an empty field. Developers already have capable open-weight models in this size class, most prominently Google’s Gemma 4 family and Alibaba’s Qwen3.6-27B — both of which position themselves around reasoning, multimodal understanding and agentic workloads. Meta’s own benchmark table compares directly against both.
Glimmer leads that three-way comparison on several agentic tests, including MCP Atlas at 75.5, DeepSearch QA at 74.6, τ³-Banking at 23.5, WildClawBench at 47.6 and GAIA2 at 43.3. It scores 51.2 on SWE-Bench Pro, versus 36.9 for Gemma4-31B and 50.2 for Qwen3.6-27B in Meta’s evaluation.
But Glimmer does not sweep the field. Qwen leads Meta’s own comparison on OSWorld-Verified (75.6 vs. Glimmer’s 65.9), TerminalBench 2.1 (60.7 vs. 51.7), SkillsBench, GDPval-AA (1141 vs. 953) and most of the multimodal benchmarks. On SWE-Bench Verified, Glimmer’s 76.0 lands just below Qwen’s 77.2. Gemma leads on GPQA Diamond and Humanity’s Last Exam.
Read honestly, the numbers make Glimmer more interesting as a specialized local-agent model than as evidence of a universal performance lead. For enterprise developers, the practical question is whether its combination of agent reliability, quantization quality, tool compatibility and decoding speed translates from benchmarks into sustained real-world workflows.
|
Model |
Developer / origin |
AA score |
Parameters / context |
Lowest tracked API price |
Access |
License |
Strongest use cases |
|
Moonshot AI; China |
60 |
2.8T total / 104B active; 1M |
$3.00 input / $15.00 output via Kimi, Fireworks or Modal (pricing) |
Custom Kimi K3 license. Large model-as-a-service operators above $20M in 12-month revenue need a separate agreement Large products may need to display “Kimi K3.” |
Frontier long-horizon coding Multimodal research and complex tool-driven agents |
||
|
Z.ai / Zhipu AI; China |
53 |
753B / 40B active; 1M |
$0.75 / $2.40 via DeepInfra FP4 (pricing) |
MIT |
Long-horizon coding and agents Million-token analysis with adjustable reasoning |
||
|
DeepSeek; China |
52 |
284B / 13B active; 1M |
$0.09 / $0.18 via DeepInfra (pricing) |
MIT |
• Extremely economical reasoning• Coding agents, terminal work and tool use |
||
|
MiniMax; China |
45 |
428B / 23B active; 1M |
$0.23 / $0.96 via CoreWeave (pricing) |
; |
MiniMax Community License. Commercial attribution required; companies above $20M yearly revenue need authorization. Includes prohibited-use conditions. |
• Native text, image and video work• Long-context coding and “cowork” agents |
|
|
Xiaomi; China |
43 |
1.02T / 42B active; 1M |
$0.35 / $0.70 via GMI (pricing) |
MIT |
Complex software engineering Agents spanning thousands of tool calls |
||
|
Thinking Machines Lab; U.S. |
42 |
975B / 41B active; 1M in weights |
$0.95 / $4.05 via DeepInfra FP8 (pricing) |
Apache 2.0 |
Customizable text, image and audio foundation Fine-tuned coding, RAG and tool-use systems |
||
|
NVIDIA; U.S. |
38 |
550B / 55B active; up to 1M in weights |
$0.37 / $1.08 via Blackbox AI (pricing) |
OpenMDW-1.1; permissive commercial and derivative-model rights |
Complex agents and long-context reasoning High-accuracy RAG, code, math and science |
||
|
Mistral AI; France |
30 |
128B dense; 256K |
$1.50 / $7.50 via Mistral (pricing) |
Modified MIT. Companies above $20M consolidated monthly revenue must obtain a commercial license or use Mistral’s service. |
Coding agents and function calling Multimodal instruction following |
||
|
Google DeepMind; U.S. |
30 |
30.7B dense; 256K |
Free on Google AI Studio’s limited tier; paid low $0.10 / $0.34 via CoreWeave (pricing) |
Apache 2.0 |
Compact multimodal reasoning and coding Manageable local or private-server deployments |
||
|
OpenAI; U.S. |
24 |
117B / 5.1B active; 131K |
$0.03 / $0.17 via CoreWeave (pricing) |
Numerous third-party APIs |
Apache 2.0 |
Reasoning structured output and tools Fine-tuning and single-80GB-GPU deployment |
|
|
Cohere; Canada |
23 |
218B / 25B active; 128K input |
Free on Cohere’s currently tracked endpoint (pricing) |
Apache 2.0 |
Enterprise RAG and grounded citations• Multilingual agents and document processing |
||
|
Meta; U.S. |
Not yet scored |
29.6B dense, including vision encoder; 131K+ |
No public metered hosted price located on launch day |
Apache 2.0 for full-precision weights, quantizations, drafter and perception encoder |
Always-on local agents on 24–32GB systems Tool use, recovery, coding and screen/document understanding |
Meta Glimmer adds to a still-small roster of genuinely open, frontier-class models from U.S. companies.
For the last two years, Chinese companies have set the pace in open source AI, with DeepSeek, Alibaba’s Qwen team, Moonshot AI’s Kimi, Zhipu’s GLM and MiniMax shipping frontier-class open models under MIT and Apache 2.0 licenses on a cadence Western labs haven’t matched.
The usage data reflects it: by May 2026, Chinese open-weight models accounted for roughly 61% of all tokens consumed on OpenRouter, with four of the five most-used models coming from Chinese labs — while Meta’s Llama, the prior open-weight leader, fell off the rankings entirely.
The U.S. counterexamples remain countable on one hand: OpenAI’s gpt-oss-120b and gpt-oss-20b, released under Apache 2.0 in August 2025 as the company’s first open weights since GPT-2; Google’s Gemma family, which is open-weight but ships under Google’s own more restrictive custom license rather than an OSI-approved one; and Thinking Machines’ Inkling.
Glimmer invites the most direct comparison to gpt-oss: both are Apache 2.0, both offer adjustable reasoning effort, and both target self-hosted deployment.
But the gpt-oss models are text-only, sparse mixture-of-experts designs built primarily for reasoning and tool use — gpt-oss-20b fits in about 16GB of memory while gpt-oss-120b targets a single 80GB data center GPU.
Glimmer stakes out different ground: a dense model with native vision input, trained end-to-end around the agent loop, shipping with its own quantized variants and speculative-decoding drafter tuned for 24GB consumer machines.
And if Zuckerberg follows through on opening Muse Spark 1.2’s weights, Meta would put an actual U.S. flagship frontier model into open circulation — something no American lab has done at that tier.
Safety remains part of the deployment architecture
Giving a local model access to tools creates a different security problem from deploying a local chatbot — and Meta’s own safety numbers show Glimmer is not uniformly stronger than its peers.
On CI Memories, a privacy benchmark where lower violation rates are better, Glimmer records 26.4 against Gemma’s 12.1 and Qwen’s 53.4. On Siren AgentDojo, a prompt-injection test, Glimmer shows a 28.4% attack-success rate versus 25.6% for Gemma and 40.3% for Qwen — while posting the highest utility score of the three at 94.2.
Meta says it evaluated Glimmer under its Advanced AI Scaling Framework and determined the model does not meet the framework’s definition of «Frontier AI» because it is generally less capable than Muse Spark. Its Preparedness Team assessed Glimmer at Moderate or lower risk across chemical/biological, cyber and loss-of-control categories — the latter two inferred from the fact that Glimmer is broadly weaker than Muse Spark 1.0, which received the same designations.
The company nevertheless recommends deploying Glimmer as part of a broader system with guardrails, including human-in-the-loop confirmation for irreversible actions. That caveat matters especially for local agents: keeping data on-device reduces exposure to cloud infrastructure, but local execution does not by itself solve prompt injection, excessive permissions or an agent taking an unintended action.
Apache 2.0 weights and a fast-growing runtime ecosystem
Meta is releasing full-precision BF16 weights, both 4-bit quantized variants, the DFlash drafter and the perception encoder — all under Apache 2.0. There is no Meta API price attached to the downloadable model, leaving total cost dependent on local hardware or whatever third-party hosting developers choose. One nuance worth noting for procurement teams: as with most «open source» model releases, it is the weights that are open — Meta has not released the training data or training code.
The broader implication is that Meta is treating the developer workstation as a credible deployment target for autonomous agents, rather than merely a place to experiment with smaller language models. Glimmer’s 30B size and 24GB target put that proposition within reach of high-end consumer hardware, while the Apache 2.0 license gives developers — and their legal departments — unusual freedom to modify and deploy it.
The next test is whether its benchmark advantages survive the messier conditions of real software repositories, enterprise tools and long-running agent sessions. If they do, the most consequential part of Glimmer may not be another set of benchmark scores — it may be that a class of agent previously expected to live behind a cloud API can increasingly live, and work, on the machine sitting under a developer’s desk.
Meta today released Muse Glimmer, a 30-billion-parameter open-weight model designed to run autonomous AI agents directly on consumer hardware — pushing agentic workloads that normally depend on cloud infrastructure onto high-end Macs and PCs.
Just as notable as what the model does is how it’s licensed. Glimmer arrives under the permissive, industry-standard Apache 2.0 open source license — the company’s first fully open release since it succeeded its open-weight Llama family in April with the proprietary Muse Spark.
In fact, Muse Glimmer launches today with a more permissive license than Llama ever carried. Llama’s bespoke community license drew years of criticism for restrictions like its 700-million-monthly-user cutoff; Apache 2.0 has no such strings, permitting unrestricted commercial use, modification and redistribution.
The weights are available on Hugging Face now. Wang said support is rolling out this week through Ollama, LM Studio, vLLM, SGLang, Together AI, Fireworks AI and OpenRouter, with optimized llama.cpp, MLX and ExecuTorch integrations landing in the coming days; Meta’s blog post also names Unsloth as a local-runtime partner and points to PyTorch’s TorchTitan for fine-tuning. The company says it is working with AMD, Arm, Dell, Intel and Nvidia to optimize performance across devices, and has published developer documentation covering custom agent scaffolds.
«Today we’re also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally,» Meta co-founder and CEO Mark Zuckerberg wrote in a post on X (under his longtime handle @finkd). «Soon we’ll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I’m proud of these releases.»
That promised Muse Spark 1.2 release would be an even bigger shift: it’s the frontier model behind Muse Code, the terminal coding agent Meta shipped just five days ago, and until today the entire Muse family was proprietary. Zuckerberg had teased at that launch that he’d «have more to share soon» on open source. Now we know what he meant.
For developers and enterprises, the practical stakes of local inference go beyond where computation happens. An agent working with files, screenshots, development environments and other sensitive context can execute those workflows without continuously sending that information to a remote inference service. Local deployment also removes network availability and per-token API charges from the inference loop — although organizations still bear hardware, electricity, deployment and management costs.
A 30B model built around the agent loop
Rather than positioning Glimmer primarily as a general chatbot, Meta trained it around the sequence of operations an autonomous agent performs: formulate a plan, call tools, interpret the results, continue working, and recover when something goes wrong.
«Just like much larger models, muse glimmer can operate as a fully capable agent via planning, tool calls, checking its own results, and failure recovery,» Alexandr Wang, Meta’s chief AI officer, wrote in a thread on X announcing the release, adding that the model «can run on 24GB of VRAM without losing agentic reliability.»
According to the model card on Hugging Face, Glimmer is a dense causal transformer with approximately 29.6 billion total parameters across 52 layers, including a dedicated ~1.8B-parameter ViT-G/14 perception encoder. It accepts interleaved text and images, produces text, supports more than 100 languages and has a stated context length of 131,072 tokens or more, with a knowledge cutoff of January 4, 2026.
That combination is intended to let an agent interpret screenshots, charts and documents while simultaneously reasoning about text and invoking external tools. Glimmer offers low, medium, high and xhigh reasoning settings — set via the system prompt — so applications can dial reasoning effort up or down per task, and Meta says it works across agentic scaffolds including OpenClaw and Hermes Agent.
The model is a distillation of Meta’s larger flagship: per the company’s technical blog post, Glimmer was pre-trained on Muse Spark’s outputs using logit distillation, mid-trained on longer-context, agent-heavy data with richer reasoning traces, then post-trained with supervised fine-tuning, on-policy distillation and reinforcement learning across general, reasoning, coding and agentic domains.
Meta demonstrated the result with a local Home Assistant workflow: in a demo video, Glimmer autonomously discovers a Home Assistant instance on the network via tool calls, queries device APIs, writes a responsive HTML/CSS/JavaScript dashboard from scratch and deploys a local server to verify its own work. That’s closer to the operational reality of enterprise agent deployments than a standalone question-answering benchmark — the model has to maintain a plan while interacting with external systems, then inspect whether its actions produced the expected result.
Compressing an agent into 24GB
The hardware story is central to the release.
At full precision, Meta says the 30B model requires more than 55GB of memory — beyond any single consumer GPU.
The company therefore developed approximately 4-bit quantized versions that shrink the language-model weights to under 20GB, leaving headroom for the pieces an operational agent also needs in memory: the KV cache, the perception encoder and a companion speculative-decoding model, all fitting within a 24GB or 32GB envelope.
In practical terms, that means the quantized builds run on consumer machines — though the upper end of them. The 24GB-targeted K-Quant-17GB configuration fits on a single high-end consumer graphics card, such as Nvidia’s RTX 3090 or RTX 4090 (both with 24GB of VRAM), while the 32GB-targeted K-Quant-Dynamic version lines up with the newer RTX 5090’s 32GB. On the Mac side, Apple Silicon’s unified memory plays the role of VRAM, so a MacBook Pro or Mac Studio with 32GB or more of memory can hold the full stack — Meta ran its own speed tests on M4 Max and M5 Max MacBook Pros. A typical 8GB or 16GB laptop, however, remains out of reach, and the full-precision BF16 release — which Meta pegs at 64GB — stays in the territory of data-center GPUs and top-spec Mac Studio configurations.
Meta reports average accuracy degradation of just 0.2% across 15 benchmarks for its K-Quant-Dynamic version targeting 32GB hardware, and 1% for the K-Quant-17GB configuration targeting 24GB hardware. Those figures are Meta’s own measurements, not independent evaluations.
Meta is also using DFlash speculative decoding to attack the other big problem with local agents: latency. Instead of generating every token sequentially, a smaller DFlash «drafter» model proposes blocks of 16 tokens that the primary model verifies in parallel, producing identical output faster.
Meta reports this raises average generation speed on an Nvidia RTX 5090 from 74.9 tokens per second to 233.4 — a 3.1x increase. An Apple M5 Max rises from 26.6 to 50.2 tokens per second (1.8x), and an M4 Max from 23.7 to 37.8 (1.5x). The tests used batch size one and greedy decoding, with Apple systems measured via ExecuTorch and the RTX 5090 via llama.cpp.
For agent applications, those multipliers matter more than they would for chat: a single user request can trigger many model turns, tool calls and verification steps, and latency accumulated at every stage can quickly make an otherwise capable agent impractical.
Glimmer enters an increasingly competitive local-model market
Meta is not entering an empty field. Developers already have capable open-weight models in this size class, most prominently Google’s Gemma 4 family and Alibaba’s Qwen3.6-27B — both of which position themselves around reasoning, multimodal understanding and agentic workloads. Meta’s own benchmark table compares directly against both.
Glimmer leads that three-way comparison on several agentic tests, including MCP Atlas at 75.5, DeepSearch QA at 74.6, τ³-Banking at 23.5, WildClawBench at 47.6 and GAIA2 at 43.3. It scores 51.2 on SWE-Bench Pro, versus 36.9 for Gemma4-31B and 50.2 for Qwen3.6-27B in Meta’s evaluation.
But Glimmer does not sweep the field. Qwen leads Meta’s own comparison on OSWorld-Verified (75.6 vs. Glimmer’s 65.9), TerminalBench 2.1 (60.7 vs. 51.7), SkillsBench, GDPval-AA (1141 vs. 953) and most of the multimodal benchmarks. On SWE-Bench Verified, Glimmer’s 76.0 lands just below Qwen’s 77.2. Gemma leads on GPQA Diamond and Humanity’s Last Exam.
Read honestly, the numbers make Glimmer more interesting as a specialized local-agent model than as evidence of a universal performance lead. For enterprise developers, the practical question is whether its combination of agent reliability, quantization quality, tool compatibility and decoding speed translates from benchmarks into sustained real-world workflows.
|
Model |
Developer / origin |
AA score |
Parameters / context |
Lowest tracked API price |
Access |
License |
Strongest use cases |
|
Moonshot AI; China |
60 |
2.8T total / 104B active; 1M |
$3.00 input / $15.00 output via Kimi, Fireworks or Modal (pricing) |
Custom Kimi K3 license. Large model-as-a-service operators above $20M in 12-month revenue need a separate agreement Large products may need to display “Kimi K3.” |
Frontier long-horizon coding Multimodal research and complex tool-driven agents |
||
|
Z.ai / Zhipu AI; China |
53 |
753B / 40B active; 1M |
$0.75 / $2.40 via DeepInfra FP4 (pricing) |
MIT |
Long-horizon coding and agents Million-token analysis with adjustable reasoning |
||
|
DeepSeek; China |
52 |
284B / 13B active; 1M |
$0.09 / $0.18 via DeepInfra (pricing) |
MIT |
• Extremely economical reasoning• Coding agents, terminal work and tool use |
||
|
MiniMax; China |
45 |
428B / 23B active; 1M |
$0.23 / $0.96 via CoreWeave (pricing) |
; |
MiniMax Community License. Commercial attribution required; companies above $20M yearly revenue need authorization. Includes prohibited-use conditions. |
• Native text, image and video work• Long-context coding and “cowork” agents |
|
|
Xiaomi; China |
43 |
1.02T / 42B active; 1M |
$0.35 / $0.70 via GMI (pricing) |
MIT |
Complex software engineering Agents spanning thousands of tool calls |
||
|
Thinking Machines Lab; U.S. |
42 |
975B / 41B active; 1M in weights |
$0.95 / $4.05 via DeepInfra FP8 (pricing) |
Apache 2.0 |
Customizable text, image and audio foundation Fine-tuned coding, RAG and tool-use systems |
||
|
NVIDIA; U.S. |
38 |
550B / 55B active; up to 1M in weights |
$0.37 / $1.08 via Blackbox AI (pricing) |
OpenMDW-1.1; permissive commercial and derivative-model rights |
Complex agents and long-context reasoning High-accuracy RAG, code, math and science |
||
|
Mistral AI; France |
30 |
128B dense; 256K |
$1.50 / $7.50 via Mistral (pricing) |
Modified MIT. Companies above $20M consolidated monthly revenue must obtain a commercial license or use Mistral’s service. |
Coding agents and function calling Multimodal instruction following |
||
|
Google DeepMind; U.S. |
30 |
30.7B dense; 256K |
Free on Google AI Studio’s limited tier; paid low $0.10 / $0.34 via CoreWeave (pricing) |
Apache 2.0 |
Compact multimodal reasoning and coding Manageable local or private-server deployments |
||
|
OpenAI; U.S. |
24 |
117B / 5.1B active; 131K |
$0.03 / $0.17 via CoreWeave (pricing) |
Numerous third-party APIs |
Apache 2.0 |
Reasoning structured output and tools Fine-tuning and single-80GB-GPU deployment |
|
|
Cohere; Canada |
23 |
218B / 25B active; 128K input |
Free on Cohere’s currently tracked endpoint (pricing) |
Apache 2.0 |
Enterprise RAG and grounded citations• Multilingual agents and document processing |
||
|
Meta; U.S. |
Not yet scored |
29.6B dense, including vision encoder; 131K+ |
No public metered hosted price located on launch day |
Apache 2.0 for full-precision weights, quantizations, drafter and perception encoder |
Always-on local agents on 24–32GB systems Tool use, recovery, coding and screen/document understanding |
Meta Glimmer adds to a still-small roster of genuinely open, frontier-class models from U.S. companies.
For the last two years, Chinese companies have set the pace in open source AI, with DeepSeek, Alibaba’s Qwen team, Moonshot AI’s Kimi, Zhipu’s GLM and MiniMax shipping frontier-class open models under MIT and Apache 2.0 licenses on a cadence Western labs haven’t matched.
The usage data reflects it: by May 2026, Chinese open-weight models accounted for roughly 61% of all tokens consumed on OpenRouter, with four of the five most-used models coming from Chinese labs — while Meta’s Llama, the prior open-weight leader, fell off the rankings entirely.
The U.S. counterexamples remain countable on one hand: OpenAI’s gpt-oss-120b and gpt-oss-20b, released under Apache 2.0 in August 2025 as the company’s first open weights since GPT-2; Google’s Gemma family, which is open-weight but ships under Google’s own more restrictive custom license rather than an OSI-approved one; and Thinking Machines’ Inkling.
Glimmer invites the most direct comparison to gpt-oss: both are Apache 2.0, both offer adjustable reasoning effort, and both target self-hosted deployment.
But the gpt-oss models are text-only, sparse mixture-of-experts designs built primarily for reasoning and tool use — gpt-oss-20b fits in about 16GB of memory while gpt-oss-120b targets a single 80GB data center GPU.
Glimmer stakes out different ground: a dense model with native vision input, trained end-to-end around the agent loop, shipping with its own quantized variants and speculative-decoding drafter tuned for 24GB consumer machines.
And if Zuckerberg follows through on opening Muse Spark 1.2’s weights, Meta would put an actual U.S. flagship frontier model into open circulation — something no American lab has done at that tier.
Safety remains part of the deployment architecture
Giving a local model access to tools creates a different security problem from deploying a local chatbot — and Meta’s own safety numbers show Glimmer is not uniformly stronger than its peers.
On CI Memories, a privacy benchmark where lower violation rates are better, Glimmer records 26.4 against Gemma’s 12.1 and Qwen’s 53.4. On Siren AgentDojo, a prompt-injection test, Glimmer shows a 28.4% attack-success rate versus 25.6% for Gemma and 40.3% for Qwen — while posting the highest utility score of the three at 94.2.
Meta says it evaluated Glimmer under its Advanced AI Scaling Framework and determined the model does not meet the framework’s definition of «Frontier AI» because it is generally less capable than Muse Spark. Its Preparedness Team assessed Glimmer at Moderate or lower risk across chemical/biological, cyber and loss-of-control categories — the latter two inferred from the fact that Glimmer is broadly weaker than Muse Spark 1.0, which received the same designations.
The company nevertheless recommends deploying Glimmer as part of a broader system with guardrails, including human-in-the-loop confirmation for irreversible actions. That caveat matters especially for local agents: keeping data on-device reduces exposure to cloud infrastructure, but local execution does not by itself solve prompt injection, excessive permissions or an agent taking an unintended action.
Apache 2.0 weights and a fast-growing runtime ecosystem
Meta is releasing full-precision BF16 weights, both 4-bit quantized variants, the DFlash drafter and the perception encoder — all under Apache 2.0. There is no Meta API price attached to the downloadable model, leaving total cost dependent on local hardware or whatever third-party hosting developers choose. One nuance worth noting for procurement teams: as with most «open source» model releases, it is the weights that are open — Meta has not released the training data or training code.
The broader implication is that Meta is treating the developer workstation as a credible deployment target for autonomous agents, rather than merely a place to experiment with smaller language models. Glimmer’s 30B size and 24GB target put that proposition within reach of high-end consumer hardware, while the Apache 2.0 license gives developers — and their legal departments — unusual freedom to modify and deploy it.
The next test is whether its benchmark advantages survive the messier conditions of real software repositories, enterprise tools and long-running agent sessions. If they do, the most consequential part of Glimmer may not be another set of benchmark scores — it may be that a class of agent previously expected to live behind a cloud API can increasingly live, and work, on the machine sitting under a developer’s desk.
Meta today released Muse Glimmer, a 30-billion-parameter open-weight model designed to run autonomous AI agents directly on consumer hardware — pushing agentic workloads that normally depend on cloud infrastructure onto high-end Macs and PCs.
Just as notable as what the model does is how it’s licensed. Glimmer arrives under the permissive, industry-standard Apache 2.0 open source license — the company’s first fully open release since it succeeded its open-weight Llama family in April with the proprietary Muse Spark.
In fact, Muse Glimmer launches today with a more permissive license than Llama ever carried. Llama’s bespoke community license drew years of criticism for restrictions like its 700-million-monthly-user cutoff; Apache 2.0 has no such strings, permitting unrestricted commercial use, modification and redistribution.
The weights are available on Hugging Face now. Wang said support is rolling out this week through Ollama, LM Studio, vLLM, SGLang, Together AI, Fireworks AI and OpenRouter, with optimized llama.cpp, MLX and ExecuTorch integrations landing in the coming days; Meta’s blog post also names Unsloth as a local-runtime partner and points to PyTorch’s TorchTitan for fine-tuning. The company says it is working with AMD, Arm, Dell, Intel and Nvidia to optimize performance across devices, and has published developer documentation covering custom agent scaffolds.
«Today we’re also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally,» Meta co-founder and CEO Mark Zuckerberg wrote in a post on X (under his longtime handle @finkd). «Soon we’ll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I’m proud of these releases.»
That promised Muse Spark 1.2 release would be an even bigger shift: it’s the frontier model behind Muse Code, the terminal coding agent Meta shipped just five days ago, and until today the entire Muse family was proprietary. Zuckerberg had teased at that launch that he’d «have more to share soon» on open source. Now we know what he meant.
For developers and enterprises, the practical stakes of local inference go beyond where computation happens. An agent working with files, screenshots, development environments and other sensitive context can execute those workflows without continuously sending that information to a remote inference service. Local deployment also removes network availability and per-token API charges from the inference loop — although organizations still bear hardware, electricity, deployment and management costs.
A 30B model built around the agent loop
Rather than positioning Glimmer primarily as a general chatbot, Meta trained it around the sequence of operations an autonomous agent performs: formulate a plan, call tools, interpret the results, continue working, and recover when something goes wrong.
«Just like much larger models, muse glimmer can operate as a fully capable agent via planning, tool calls, checking its own results, and failure recovery,» Alexandr Wang, Meta’s chief AI officer, wrote in a thread on X announcing the release, adding that the model «can run on 24GB of VRAM without losing agentic reliability.»
According to the model card on Hugging Face, Glimmer is a dense causal transformer with approximately 29.6 billion total parameters across 52 layers, including a dedicated ~1.8B-parameter ViT-G/14 perception encoder. It accepts interleaved text and images, produces text, supports more than 100 languages and has a stated context length of 131,072 tokens or more, with a knowledge cutoff of January 4, 2026.
That combination is intended to let an agent interpret screenshots, charts and documents while simultaneously reasoning about text and invoking external tools. Glimmer offers low, medium, high and xhigh reasoning settings — set via the system prompt — so applications can dial reasoning effort up or down per task, and Meta says it works across agentic scaffolds including OpenClaw and Hermes Agent.
The model is a distillation of Meta’s larger flagship: per the company’s technical blog post, Glimmer was pre-trained on Muse Spark’s outputs using logit distillation, mid-trained on longer-context, agent-heavy data with richer reasoning traces, then post-trained with supervised fine-tuning, on-policy distillation and reinforcement learning across general, reasoning, coding and agentic domains.
Meta demonstrated the result with a local Home Assistant workflow: in a demo video, Glimmer autonomously discovers a Home Assistant instance on the network via tool calls, queries device APIs, writes a responsive HTML/CSS/JavaScript dashboard from scratch and deploys a local server to verify its own work. That’s closer to the operational reality of enterprise agent deployments than a standalone question-answering benchmark — the model has to maintain a plan while interacting with external systems, then inspect whether its actions produced the expected result.
Compressing an agent into 24GB
The hardware story is central to the release.
At full precision, Meta says the 30B model requires more than 55GB of memory — beyond any single consumer GPU.
The company therefore developed approximately 4-bit quantized versions that shrink the language-model weights to under 20GB, leaving headroom for the pieces an operational agent also needs in memory: the KV cache, the perception encoder and a companion speculative-decoding model, all fitting within a 24GB or 32GB envelope.
In practical terms, that means the quantized builds run on consumer machines — though the upper end of them. The 24GB-targeted K-Quant-17GB configuration fits on a single high-end consumer graphics card, such as Nvidia’s RTX 3090 or RTX 4090 (both with 24GB of VRAM), while the 32GB-targeted K-Quant-Dynamic version lines up with the newer RTX 5090’s 32GB. On the Mac side, Apple Silicon’s unified memory plays the role of VRAM, so a MacBook Pro or Mac Studio with 32GB or more of memory can hold the full stack — Meta ran its own speed tests on M4 Max and M5 Max MacBook Pros. A typical 8GB or 16GB laptop, however, remains out of reach, and the full-precision BF16 release — which Meta pegs at 64GB — stays in the territory of data-center GPUs and top-spec Mac Studio configurations.
Meta reports average accuracy degradation of just 0.2% across 15 benchmarks for its K-Quant-Dynamic version targeting 32GB hardware, and 1% for the K-Quant-17GB configuration targeting 24GB hardware. Those figures are Meta’s own measurements, not independent evaluations.
Meta is also using DFlash speculative decoding to attack the other big problem with local agents: latency. Instead of generating every token sequentially, a smaller DFlash «drafter» model proposes blocks of 16 tokens that the primary model verifies in parallel, producing identical output faster.
Meta reports this raises average generation speed on an Nvidia RTX 5090 from 74.9 tokens per second to 233.4 — a 3.1x increase. An Apple M5 Max rises from 26.6 to 50.2 tokens per second (1.8x), and an M4 Max from 23.7 to 37.8 (1.5x). The tests used batch size one and greedy decoding, with Apple systems measured via ExecuTorch and the RTX 5090 via llama.cpp.
For agent applications, those multipliers matter more than they would for chat: a single user request can trigger many model turns, tool calls and verification steps, and latency accumulated at every stage can quickly make an otherwise capable agent impractical.
Glimmer enters an increasingly competitive local-model market
Meta is not entering an empty field. Developers already have capable open-weight models in this size class, most prominently Google’s Gemma 4 family and Alibaba’s Qwen3.6-27B — both of which position themselves around reasoning, multimodal understanding and agentic workloads. Meta’s own benchmark table compares directly against both.
Glimmer leads that three-way comparison on several agentic tests, including MCP Atlas at 75.5, DeepSearch QA at 74.6, τ³-Banking at 23.5, WildClawBench at 47.6 and GAIA2 at 43.3. It scores 51.2 on SWE-Bench Pro, versus 36.9 for Gemma4-31B and 50.2 for Qwen3.6-27B in Meta’s evaluation.
But Glimmer does not sweep the field. Qwen leads Meta’s own comparison on OSWorld-Verified (75.6 vs. Glimmer’s 65.9), TerminalBench 2.1 (60.7 vs. 51.7), SkillsBench, GDPval-AA (1141 vs. 953) and most of the multimodal benchmarks. On SWE-Bench Verified, Glimmer’s 76.0 lands just below Qwen’s 77.2. Gemma leads on GPQA Diamond and Humanity’s Last Exam.
Read honestly, the numbers make Glimmer more interesting as a specialized local-agent model than as evidence of a universal performance lead. For enterprise developers, the practical question is whether its combination of agent reliability, quantization quality, tool compatibility and decoding speed translates from benchmarks into sustained real-world workflows.
|
Model |
Developer / origin |
AA score |
Parameters / context |
Lowest tracked API price |
Access |
License |
Strongest use cases |
|
Moonshot AI; China |
60 |
2.8T total / 104B active; 1M |
$3.00 input / $15.00 output via Kimi, Fireworks or Modal (pricing) |
Custom Kimi K3 license. Large model-as-a-service operators above $20M in 12-month revenue need a separate agreement Large products may need to display “Kimi K3.” |
Frontier long-horizon coding Multimodal research and complex tool-driven agents |
||
|
Z.ai / Zhipu AI; China |
53 |
753B / 40B active; 1M |
$0.75 / $2.40 via DeepInfra FP4 (pricing) |
MIT |
Long-horizon coding and agents Million-token analysis with adjustable reasoning |
||
|
DeepSeek; China |
52 |
284B / 13B active; 1M |
$0.09 / $0.18 via DeepInfra (pricing) |
MIT |
• Extremely economical reasoning• Coding agents, terminal work and tool use |
||
|
MiniMax; China |
45 |
428B / 23B active; 1M |
$0.23 / $0.96 via CoreWeave (pricing) |
; |
MiniMax Community License. Commercial attribution required; companies above $20M yearly revenue need authorization. Includes prohibited-use conditions. |
• Native text, image and video work• Long-context coding and “cowork” agents |
|
|
Xiaomi; China |
43 |
1.02T / 42B active; 1M |
$0.35 / $0.70 via GMI (pricing) |
MIT |
Complex software engineering Agents spanning thousands of tool calls |
||
|
Thinking Machines Lab; U.S. |
42 |
975B / 41B active; 1M in weights |
$0.95 / $4.05 via DeepInfra FP8 (pricing) |
Apache 2.0 |
Customizable text, image and audio foundation Fine-tuned coding, RAG and tool-use systems |
||
|
NVIDIA; U.S. |
38 |
550B / 55B active; up to 1M in weights |
$0.37 / $1.08 via Blackbox AI (pricing) |
OpenMDW-1.1; permissive commercial and derivative-model rights |
Complex agents and long-context reasoning High-accuracy RAG, code, math and science |
||
|
Mistral AI; France |
30 |
128B dense; 256K |
$1.50 / $7.50 via Mistral (pricing) |
Modified MIT. Companies above $20M consolidated monthly revenue must obtain a commercial license or use Mistral’s service. |
Coding agents and function calling Multimodal instruction following |
||
|
Google DeepMind; U.S. |
30 |
30.7B dense; 256K |
Free on Google AI Studio’s limited tier; paid low $0.10 / $0.34 via CoreWeave (pricing) |
Apache 2.0 |
Compact multimodal reasoning and coding Manageable local or private-server deployments |
||
|
OpenAI; U.S. |
24 |
117B / 5.1B active; 131K |
$0.03 / $0.17 via CoreWeave (pricing) |
Numerous third-party APIs |
Apache 2.0 |
Reasoning structured output and tools Fine-tuning and single-80GB-GPU deployment |
|
|
Cohere; Canada |
23 |
218B / 25B active; 128K input |
Free on Cohere’s currently tracked endpoint (pricing) |
Apache 2.0 |
Enterprise RAG and grounded citations• Multilingual agents and document processing |
||
|
Meta; U.S. |
Not yet scored |
29.6B dense, including vision encoder; 131K+ |
No public metered hosted price located on launch day |
Apache 2.0 for full-precision weights, quantizations, drafter and perception encoder |
Always-on local agents on 24–32GB systems Tool use, recovery, coding and screen/document understanding |
Meta Glimmer adds to a still-small roster of genuinely open, frontier-class models from U.S. companies.
For the last two years, Chinese companies have set the pace in open source AI, with DeepSeek, Alibaba’s Qwen team, Moonshot AI’s Kimi, Zhipu’s GLM and MiniMax shipping frontier-class open models under MIT and Apache 2.0 licenses on a cadence Western labs haven’t matched.
The usage data reflects it: by May 2026, Chinese open-weight models accounted for roughly 61% of all tokens consumed on OpenRouter, with four of the five most-used models coming from Chinese labs — while Meta’s Llama, the prior open-weight leader, fell off the rankings entirely.
The U.S. counterexamples remain countable on one hand: OpenAI’s gpt-oss-120b and gpt-oss-20b, released under Apache 2.0 in August 2025 as the company’s first open weights since GPT-2; Google’s Gemma family, which is open-weight but ships under Google’s own more restrictive custom license rather than an OSI-approved one; and Thinking Machines’ Inkling.
Glimmer invites the most direct comparison to gpt-oss: both are Apache 2.0, both offer adjustable reasoning effort, and both target self-hosted deployment.
But the gpt-oss models are text-only, sparse mixture-of-experts designs built primarily for reasoning and tool use — gpt-oss-20b fits in about 16GB of memory while gpt-oss-120b targets a single 80GB data center GPU.
Glimmer stakes out different ground: a dense model with native vision input, trained end-to-end around the agent loop, shipping with its own quantized variants and speculative-decoding drafter tuned for 24GB consumer machines.
And if Zuckerberg follows through on opening Muse Spark 1.2’s weights, Meta would put an actual U.S. flagship frontier model into open circulation — something no American lab has done at that tier.
Safety remains part of the deployment architecture
Giving a local model access to tools creates a different security problem from deploying a local chatbot — and Meta’s own safety numbers show Glimmer is not uniformly stronger than its peers.
On CI Memories, a privacy benchmark where lower violation rates are better, Glimmer records 26.4 against Gemma’s 12.1 and Qwen’s 53.4. On Siren AgentDojo, a prompt-injection test, Glimmer shows a 28.4% attack-success rate versus 25.6% for Gemma and 40.3% for Qwen — while posting the highest utility score of the three at 94.2.
Meta says it evaluated Glimmer under its Advanced AI Scaling Framework and determined the model does not meet the framework’s definition of «Frontier AI» because it is generally less capable than Muse Spark. Its Preparedness Team assessed Glimmer at Moderate or lower risk across chemical/biological, cyber and loss-of-control categories — the latter two inferred from the fact that Glimmer is broadly weaker than Muse Spark 1.0, which received the same designations.
The company nevertheless recommends deploying Glimmer as part of a broader system with guardrails, including human-in-the-loop confirmation for irreversible actions. That caveat matters especially for local agents: keeping data on-device reduces exposure to cloud infrastructure, but local execution does not by itself solve prompt injection, excessive permissions or an agent taking an unintended action.
Apache 2.0 weights and a fast-growing runtime ecosystem
Meta is releasing full-precision BF16 weights, both 4-bit quantized variants, the DFlash drafter and the perception encoder — all under Apache 2.0. There is no Meta API price attached to the downloadable model, leaving total cost dependent on local hardware or whatever third-party hosting developers choose. One nuance worth noting for procurement teams: as with most «open source» model releases, it is the weights that are open — Meta has not released the training data or training code.
The broader implication is that Meta is treating the developer workstation as a credible deployment target for autonomous agents, rather than merely a place to experiment with smaller language models. Glimmer’s 30B size and 24GB target put that proposition within reach of high-end consumer hardware, while the Apache 2.0 license gives developers — and their legal departments — unusual freedom to modify and deploy it.
The next test is whether its benchmark advantages survive the messier conditions of real software repositories, enterprise tools and long-running agent sessions. If they do, the most consequential part of Glimmer may not be another set of benchmark scores — it may be that a class of agent previously expected to live behind a cloud API can increasingly live, and work, on the machine sitting under a developer’s desk.
