The LLM benchmark aggregator.

Muse Glimmer-30B

Meta · 30B · open weights

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Agents51.1−17—/267✱0/5

Tool use52.8−21.2—/81✱0/1

Core abilities53.9−13.5—/204✱0/3

Long context64.3−0.3—/0✱✱

Reasoning49.2−17.7—/178✱0/4

Science50−13.5—/122✱0/1

Expert reasoning44.8−16.1—/0✱✱

Coding51.3−17.8—/164✱0/5

Agentic coding51.2−17.5—/157✱0/4

Code generation54.4−15.6—/77✱0/2

Knowledge46.7−16.8—/138✱0/2

Factuality45−14—/0✱✱

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Muse Glimmer-30B, left for the other.

NVIDIA Nemotron 3 Nano 4Bwon 8 · lost 0

GPT-4o Miniwon 8 · lost 0

GPT-4.1 Nanowon 8 · lost 0

Granite 4.2 3Bwon 9 · lost 0

Nemotron 3 Superwon 12 · lost 0

GPT-4.1 Miniwon 7 · lost 1

Llama 4 Maverick Instructwon 7 · lost 1

Granite 4.2 30Bwon 10 · lost 2

G9v3 39A5Bwon 10 · lost 2

Gemma 4 31Bwon 10 · lost 2

Granite 4.2 8Bwon 7 · lost 2

Qwen3.5 9Bwon 7 · lost 2

G9v3 3Bwon 6 · lost 2

Grok 4.3won 6 · lost 2

Gemini 3.5 Flash Litewon 6 · lost 2

GPT-5.4 Miniwon 6 · lost 2

GPT-5.4 Nanowon 6 · lost 2

Command Awon 6 · lost 2

Gemma 4 12Bwon 6 · lost 2

DeepSeek V3won 6 · lost 2

Qwen3.6 35B A3Bwon 9 · lost 3

K2 Horizon 32Bwon 9 · lost 3

K2 Horizon 3.7Bwon 6 · lost 3

Qwen3.5 4Bwon 5 · lost 3

Qwen3.6 Pluswon 5 · lost 3

Kimi K2.6won 5 · lost 3

Claude Haiku 4.5won 5 · lost 3

Inklingwon 7 · lost 5

Nemotron 3 Ultrawon 7 · lost 5

K2 Horizon 7Bwon 5 · lost 4

Claude Sonnet 4.6won 4 · lost 4

GLM-4.6won 4 · lost 4

Gemini 3.6 Flashwon 4 · lost 4

Qwen3.7 Maxwon 4 · lost 4

Gemini 3.5 Flashwon 4 · lost 4

Claude Sonnet 4won 4 · lost 4

Gemini 3.1 Prowon 4 · lost 4

MiniMax M3won 6 · lost 6

GPT-5.2won 4 · lost 5

K2 Horizon MoVA 36B A4Bwon 5 · lost 7

GLM-5.3 Flashwon 3 · lost 5

Kimi K2won 3 · lost 5

DeepSeek V3.2won 3 · lost 5

GPT-4.1won 3 · lost 5

Claude Sonnet 4.5won 3 · lost 5

Gemini 2.5 Prowon 3 · lost 5

GPT-5 Miniwon 3 · lost 5

DeepSeek V4 Flashwon 3 · lost 6

Gemini 2.5 Flashwon 3 · lost 6

Claude Opus 4.5won 3 · lost 6

Qwen3.8 27Bwon 4 · lost 8

Muse Spark 1.1won 2 · lost 6

Qwen3 235B A22B Instructwon 2 · lost 6

Grok 4.5won 2 · lost 6

Gemini 3 Prowon 2 · lost 6

K2 Horizon 375B A23Bwon 3 · lost 9

GLM-5.3won 2 · lost 7

O4 Miniwon 2 · lost 7

GPT-5.6 Lunawon 2 · lost 10

GLM-5.2won 2 · lost 10

Claude Sonnet 5won 2 · lost 10

GPT-5.6 Terrawon 2 · lost 10

Gemini 3.8 Flashwon 1 · lost 7

GPT-5.4won 1 · lost 7

Gemini 3.7 Flashwon 1 · lost 7

Grok 4.6won 1 · lost 7

Qwen3.8 Maxwon 1 · lost 7

DeepSeek V4 Prowon 1 · lost 8

DeepSeek V4.1 Flashwon 1 · lost 8

Claude Opus 4.8won 0 · lost 8

Claude Opus 4.7won 0 · lost 8

Grok 4won 0 · lost 8

GPT-5.5won 0 · lost 8

Claude Fable 5.1won 0 · lost 8

Claude Fable 5won 0 · lost 8

O3won 0 · lost 9

Claude Opus 5won 0 · lost 9

Kimi K3won 0 · lost 9

GPT-5.6 Solwon 0 · lost 9

§ 3 · Sources

Where the numbers come from

1 publication, 9 figures. Every one links to the page it was read from.

✱✱ IFM — K2 Horizon launch ↗the model’s own publisher9 measuresread 2026-09-27

tau3-Banking 23.5%Terminal-Bench 2.1 51.7%SciCode 43.6%Humanity's Last Exam (without tools) 22%GPQA Diamond 83.5%CritPt 2.6%AA-LCR 80%AA-Omniscience Accuracy 27%AA-Omniscience Non-Hallucination 18.1%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=muse-glimmer-30b)](https://publicai.io/model-index/m/muse-glimmer-30b)