GPQA DiamondVerified
83.5%
GPQA Diamond (AA run)
Models / Meta
Muse Glimmer
Muse · Released 2026-08-10
Meta's 30B open-weight local-agent model (10 Aug 2026). AA hosted list price $0.325 / $1.35 per 1M tokens; weights are Apache 2.0.
On-device Muse variant for always-on local agents, function calling, and local coding. Prefer sourced ledger rows over launch claims; AA scores sync from the public snapshot.
Composite56
ContextUnlisted
Input / 1M$0.33
Output / 1M$1.4
Knowledge cutoffunpublished
Output speed131 t/s
Statussolid
Family profile
Best published score in each covered benchmark family.
66/100 avg
- Reasoning
- 84
- Coding
- 45
- Agentic
- 52
- Long context
- 83
4 tested benchmark families
Editor's note
Editorial depth still needs a desk pass — scores below are ledger-backed, not hands-on claims. This is the local 30B Muse lane, not a Spark replacement.Benchmark placements
Per-benchmark positions
Where Muse Glimmer places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.
| Benchmark | Family | Rank | Score | Source | Date |
|---|---|---|---|---|---|
| AA-LCR | Long context | #12/ 184 | 83 | Artificial Analysis | 2026-09-28 |
| AA output speed | Performance | #37/ 195 | Artificial Analysis | 2026-10-12 | |
| SciCode | Coding | #66/ 97 | 45 | Artificial Analysis | 2026-09-28 |
| Terminal-Bench | Agentic | #72/ 167 | 52 | Artificial Analysis | 2026-09-28 |
| GPQA Diamond | Reasoning | #116/ 184 | 84 | Artificial Analysis | 2026-09-28 |
| Humanity's Last Exam | Reasoning | #123/ 186 | 22 | Artificial Analysis | 2026-09-28 |
Family context
Strongest rows in each family
The three highest-scoring models with pages in each capability family. Where Muse Glimmer shows up, it's highlighted.
ReasoningMuse Glimmer · 84
- 196
- 295
- 395
Math
- 152
- 252
- 350
CodingMuse Glimmer · 45
- 174
- 273
- 370
AgenticMuse Glimmer · 52
- 191
- 289
- 389
Long contextMuse Glimmer · 83
- 189
- 286
- 385
Newest receipts
6 recentLast 120 days
Benchmark rows added to the public ledger for Muse Glimmer in the last 120 days. Older rows live in the full table below.
| Benchmark | Family | Score | Source | Days ago |
|---|---|---|---|---|
| AA output speed | Performance | 131 t/s | Artificial Analysis | 10 |
| AA-LCR | Long context | 83.33% | Artificial Analysis | 24 |
| GPQA Diamond | Reasoning | 83.5% | Artificial Analysis | 24 |
| Humanity's Last Exam | Reasoning | 22% | Artificial Analysis | 24 |
| SciCode | Coding | 44.9% | Artificial Analysis | 24 |
| Terminal-Bench | Agentic | 51.69% | Artificial Analysis | 24 |
Family coverage
Best score per family
A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.
Price vs. performance
Where Muse Glimmer sits on the scatter
Hover any dot for name, score, and input price. Keyboard: tab through the top twelve, or use the ranking below.
AnthropicOpenAIGoogleAlibabaMetaZhipuxAIDeepSeekMiniMaxXiaomiMoonshotOtherMeituanMistralNVIDIACohere
Full benchmark ledger
Every catalog benchmark for Muse Glimmer. Scores link to the original source; gaps mean no public row exists yet.
6 of 13 catalog benchmarks have a sourced row for Muse Glimmer.
Reasoning
Humanity's Last ExamVerified
22%
HLE (AA run)
Coding
SciCodeVerified
44.9%
SciCode (AA run)
Agentic
Terminal-BenchVerified
51.69%
Terminal-Bench v2.1 (AA run)
Long context
AA-LCRVerified
83.33%
AA-LCR (AA run)
Performance
AA output speedVerified
131 t/s
Median output tokens/s (1k prompt, default provider)
Changelog
Version history
- Muse Glimmer open-weighted (Apache 2.0, 30B).
Sources
Where the numbers come from
- Meta Muse Glimmer announcement· checked 2026-08-13
- Hugging Face Muse-Glimmer-30B· checked 2026-08-13
- Artificial Analysis Muse Glimmer· checked 2026-08-13
Questions this model page answers
Every answer below is assembled from the dated fields on this page. Nothing is written separately for search.
What is Muse Glimmer?
Muse Glimmer is a Meta model: Meta's 30B open-weight local-agent model (10 Aug 2026). AA hosted list price $0.325 / $1.35 per 1M tokens; weights are Apache 2.0.
What is Muse Glimmer for?
On-device Muse variant for always-on local agents, function calling, and local coding. Prefer sourced ledger rows over launch claims; AA scores sync from the public snapshot.
What does Muse Glimmer cost?
Muse Glimmer lists $0.33 input and $1.4 output per 1M tokens. Pricing last checked 2026-08-13.
How large is the Muse Glimmer context window?
Muse Glimmer does not publish a context window on the sources checked for this page.
What is the VerdictPal status of Muse Glimmer?
Muse Glimmer carries Solid status on the model atlas.
The Pack · Editorial newsletter
Know when the rules move.
One short email when a university changes its AI rule for theses, or a research tool changes its pricing or privacy terms. No tracking, no third-party analytics. Unsubscribe in one click.