Muse Glimmer 30B
meta/muse-glimmer-30b- Meta
- Family · Muse Glimmer
- Open weights
- Text · Image → Text
- Reasoning · adjustable
- Tools · JSON · Vision
Not on OSMEXO yet
Spec & list price
Context128K
Max output16K
Input / 1M$0.30
Output / 1M$1.204× input
Usage
Tokens / day1.1B▲ 27% 7d
Usage rank#194
Requests / day336.1K
Tool calls / req0.39
Per request · 30 days
Avg input8.6K0.6× median
Avg output5100.7× median
Cache hit79%1.1× median
Reasoning share75%1.3× median
Tool-call failures2.6%1.4× median
Consumption
Peak 12.9B 7 OctLow 508.5M 10 OctRank today #194
News & trends
4 mentions this week ▼ 50%this model · videos, posts, articles · 30 days
Trending · 14 days
- Muse-Glimmer-30B FP8 – 110 t/s combined decode – 2x Arc Pro B70 (TP2, dflash d15, 180W)1 upvotes
- Muse Glimmer 30B on M5 Pro Mac mini 48GB FP4 vs M5 Ultra Mac Studio 256GB FP8509 views
- Muse Glimmer 30B spends a third of its output thinking. Do rules change that?163 views
News · latest from Meta
- Meta Networking at the 2026 OCP Global Summit: Open Innovation for Scale-Up, Scale-Out, and Scale-Across AI NetworksMeta Engineering · 26 Oct
- Scale Agentic Coding Workflows with Meta’s Muse Spark Open on NVIDIA GB300 NVL72NVIDIA · 15 Oct
- Meta Muse Spark Open is live on Together AITogether AI · 15 Oct
- Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta15 Oct
Speed
Output speed#96/33491 tok/s1.4× faster than median
First token#490.47 s2.2× quicker than median
Technical details
- Reasoning
- on · medium
- Tokenizer
- Other
- License
- Apache 2.0
- Parameters
- 29.8B
- Moderation
- none
- Downloads · likes
- 194.1K · 2.1K
Reasoning effort
- xhigh
- high
- medium · default
- low
Request parameters 18
- tools
- tool_choice
- response_format
- structured_outputs
- reasoning
- reasoning_effort
- include_reasoning
- temperature
- top_p
- top_k
+8 more
- min_p
- max_tokens
- stop
- seed
- frequency_penalty
- logit_bias
- presence_penalty
- repetition_penalty
When to use
Best for
- Self-hosted image-aware chat (open weights)
- Reasoning tasks with adjustable effort
- Document and screenshot Q&A
- Tool calling with structured output
Think twice for
- Very large outputs (short output cap)
- Very long documents (limited context)
- Audio or video input
Research papers
- Muse Glimmer 30B: An On-Device Agentic Model17 Oct
- MetaEncoder: Exploring the Limit of Bi-Encoders for Multimodal System One Decision Making with Natural Language Interface8 Oct
- Curating Merchant-Matching Training Data with Two Confidence-Gated Local LLM Judges27 Sep
- MetroLLM-Bench: Evaluating Language Models as Transit Kiosk Runtimes9 Sep
Social
Most watched and discussed · 90 days
- Muse Glimmer 30B on M5 Pro Mac mini 48GB FP4 vs M5 Ultra Mac Studio 256GB FP8Elite Test Engineering · 509 views · 26 Oct
- Muse Glimmer 30B spends a third of its output thinking. Do rules change that?PC Magik Lab · 163 views · 22 Oct
- Meta最新研究：Muse Glimmer 30B 挑戰邊緣設備端AI代理與工具調用新極限NickBrainEvo · 136 views · 21 Oct
- Muse Glimmer 30B: An On-Device Agentic Model (Oct 2026)AI Paper Slop · 107 views · 25 Oct
- Muse Glimmer 30B: An Agentic Model for Consumer HardwareAI Paper Slop · 45 views · 25 Oct
- Podcast \ How Metas Muse Glimmer 30B WorksVinh Nguyen · 41 views · 23 Oct
- How Meta's Muse Glimmer 30B WorksVinh Nguyen · 30 views · 23 Oct
- Meta Muse Glimmer 30B: Free AI Visual Coding ModelSanjeev Kumar · 29 views · 21 Oct
- Squeezing an Autonomous Agent into 32GB: Inside Meta’s Muse Glimmer 30BThe Classical Remix · 28 views · 22 Oct
- How Muse Glimmer 30B Really Works: On-Device Agent, KV Cache, Distillation & Quantization (2.4h)PARTH VADERA · 23 views · 2 Nov
- Meta Muse Glimmer 30B: Free AI Visual Coding ModelSanjeev Kumar · 11 views · 31 Oct
- Meta Muse Glimmer 30B: Run It on Your Own GPUDX Today Podcast · 6 views · 28 Oct
Same family
Muse Glimmer- 17.5
Want Muse Glimmer 30B on OSMEXO?
Sign in to request it. Requests are reviewed; a request does not guarantee availability.