Back to models
Find GGUF downloads (opens in a new tab)Original weights (opens in a new tab)Access approval required
Intelligence Index
Extra high setting
38.9#24 of 83measured models
Measured · Artificial Analysis v4.3.2- API price from
- $1.25 in / $4.25 out / 1M tokOpenRouter · BaseTen · FP4
- Context window
- 1M tok
Performance profile
ReasoningRank #27 of 107
89.0%
GPQA DiamondCodingRank #13 of 94
58.4%
SciCodeCost efficiencyRank #40 of 67
46.1pts / $
Intelligence ÷ eval costCapabilities
Context window
Modalities
Measured setting
Integration
Reasoning trade-offs
Source (opens in a new tab)Muse Spark Open
Cost per task (USD)
Muse Spark OpenXhigh
- Intelligence Index
- 38.9
- Cost / task
- $0.8435
- Evaluation time
- 5.2m
Pricing & access
USD per 1M tokens
- Input
- $1.25
- Output
- $4.25
- Cached input
- $0.15
Run locally
- Original download
- 1,644.3 GB
- App memory
- Unverified
- Documented software
- MLX / SGLang / vLLMSetup guide (opens in a new tab)
Find GGUF downloads (opens in a new tab)Original weights (opens in a new tab)Access approval requiredBenchmarks
Full leaderboardSource (opens in a new tab)43.9%, rank 16 of 12658.4%, rank 13 of 948.1%, rank 30 of 8778.3%, rank 41 of 1251,401 Elo, rank 31 of 101
Humanity’s Last Exam
Frontier reasoning- 1Claude Opus 5.5Max61.4%
- 2Claude Fable 5.1Max59.1%
- 3Gemini 4 ArgonHigh57.1%
- 13Kimi-K3Max46.9%
- 14Step 5 PreviewReasoning46.5%
- 15Claude Haiku 5.5Max44.4%
- 16Muse Spark OpenExtra high43.9%
- 17JT-4.1 Flash 236B A21BReasoning43.6%
- 18Grok 4.7Extra high43.1%
- 19Qwen3.8 Max (0902)Reasoning43.1%
Peers use their best measured result on this benchmark.
Alternatives
- Scores higher
- Intelligence Index
- 39.5+0.6
- Cost / task
- $0.265−69%
- Higher score, lower costMiMo-V2.6-ProReasoning
- Intelligence Index
- 46.3+7.5
- Cost / task
- $0.133−84%
- More from Meta
- Intelligence Index
- 48.1+9.2
- Cost / task
- $1.60+90%
Each model at its highest-scoring measured setting · Intelligence Index v4.3.2 and its evaluation cost per task.