Skip to content
Model Pareto

Muse Glimmer

MetaOpen weightsresearch.meta.ai ↗
Compare with nearest rivalsCompare with…
Input
$0.325/M tok
Output
$1.35/M tok
Blended
$0.581/M tok · 3:1
Speed
133tok/s
Context
131Ktokens

Open weights (30B, Apache 2.0); no first-party API, price is the Artificial Analysis third-party host price.

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs, compared with the API price above.

Hardware
1× B200 180GB
BF16 weights
Throughput
~4,025 tok/s
many requests batched
Price
$0.23–2.05 /M tok
busy → light use
On your own machine
RTX 5090 32GB (int4, ~85 tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • BF16 weights (29.6B params) + 25% KV-cache headroom ≈ 74 GB
  • 1× B200 180GB at $5.98–$14.24/GPU-hour on-demand (2026-09-24)
  • ~4,025 output tok/s aggregate at batch 115 (bandwidth-bound)
  • Blended 3:1 input:output; prefill ~17,958 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Coding

DeepSWE v1.1Estimated
~#28 of 31 (estimated)~50.9%
  1. #1DeepSeek V4.1 Flash74.2%
  2. #2Grok 4.772.6%
  3. #27Qwen3.8 Max (0902)51.0%
  4. ~#28Muse GlimmerEstimated50.9%
  5. #28Qwen3.8 27B42.2%
No published score — estimated from related benchmarks
LiveCodeBenchEstimated
~#8 of 20 (estimated)~70.9%
  1. #1Qwen3.8-Omni-Flash92.6%
  2. #2Qwen3.8 Flash-Next91.9%
  3. #7gpt-oss-20b77.7%
  4. ~#8Muse GlimmerEstimated70.9%
  5. #8Mistral Small 464.0%
No published score — estimated from related benchmarks
AA Coding Agent IndexEstimated
~last of 15 (estimated)below measured range (<41.1)
  1. #1Claude Opus 5.566.0
  2. #2Claude Fable 5.162.2
  3. #14GPT-6 Luna41.1
  4. ~#15Muse GlimmerEstimated<41.1
No published score — estimated from related benchmarks

Writing & chat

MMMLUEstimated
~#3 of 5 (estimated)~86.6%
  1. #1Gemini 3.1 Pro92.6%
  2. #2Gemma 4 31B88.4%
  3. ~#3Muse GlimmerEstimated86.6%
  4. #3Claude Haiku 4.583.0%
No published score — estimated from related benchmarks
LMArena Text, Non-English (Elo)Few votesVerified
#32 of 531418 Elo
  1. #1Claude Opus 5.51501
  2. #2Claude Fable 5.1Few votes1496
  3. #31Gemma 4 26B A4BFew votes1423
  4. #32Muse GlimmerFew votes1418
  5. #33MiMo-V2.51411
lmarena.ai · 2026-09-25
LMArena Text (Elo)Few votesVerified
#35 of 531424 Elo
  1. #1Claude Opus 5.51509
  2. #2Claude Fable 5.11501
  3. #33Nemotron 3 Ultra1426
  4. #35Muse GlimmerFew votes1424
  5. #36Qwen3.5 122B A10B1417
lmarena.ai · 2026-09-25

Vision

MMMUEstimated
~#2 of 3 (estimated)~71.3%
  1. #1Claude Haiku 4.573.2%
  2. ~#2Muse GlimmerEstimated71.3%
  3. #2Llama 4 Maverick69.4%
No published score — estimated from related benchmarks

Hard reasoning

Agents

τ²-bench (Telecom)Estimated
~#38 of 116 (estimated)~70.2%
  1. #1Step 3.7 Flash98.5%
  2. #2Gemini 3.1 Pro95.6%
  3. #37Nova 2.0 Lite72.8%
  4. ~#38Muse GlimmerEstimated70.2%
  5. #38Apriel-v1.6-15B-Thinker69.3%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~last of 5 (estimated)below measured range (<47.5%)
  1. #1GLM-5.3 Flash59.1%
  2. #2Kimi K358.3%
  3. #4GPT-5.547.5%
  4. ~#5Muse GlimmerEstimated<47.5%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#15 of 18 (estimated)~66.3%
  1. #1Atria Dawn Preview92.5%
  2. #2GPT-6 Astra91.5%
  3. #14Inkling77.1%
  4. ~#15Muse GlimmerEstimated66.3%
  5. #15Mistral Medium 3.548.6%
No published score — estimated from related benchmarks