Back to models
Intelligence Index
Reasoning setting
29.2#35 of 83measured models
Measured · Artificial Analysis v4.3.2- API price from
- $0.30 in / $1.20 out / 1M tokOpenRouter · Parasail · FP8
- Context window
- 1M tok
Performance profile
ReasoningRank #8 of 107
92.9%
GPQA DiamondCodingRank #42 of 94
47.1%
SciCodeCost efficiencyRank #32 of 67
57.6pts / $
Intelligence ÷ eval costCapabilities
Context window
Modalities
Measured setting
Integration
Reasoning trade-offs
Source (opens in a new tab)MiniMax-M3
Cost per task (USD) · log scale
MiniMax-M3Reasoning
- Intelligence Index
- 29.2
- Cost / task
- $0.5076
- Evaluation time
- 8.8m
Pricing & access
USD per 1M tokens
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.06
Run locally
GGUF
Context
Estimated memory
~271.5 GiB
RAM + VRAM
Estimate assumptions
One text chat at the selected context, with a 16-bit cache and app allowance. Operating system and image processing are extra; optimized caches can use less.
384 GiB RAMwithout GPU offload
Memory budgetOne text chat
- Weights
- 245.9 GiB
- Context cache
- 1 GiB16-bit cache
- App allowance
- 24.6 GiB
- Download
- 264 GB7 files · keep together
- Disk space
- ~296 GB
Benchmarks
Full leaderboardSource (opens in a new tab)39.0%, rank 28 of 12647.1%, rank 42 of 942.0%, rank 37 of 8783.0%, rank 13 of 1251,242 Elo, rank 35 of 101
Humanity’s Last Exam
Frontier reasoning- 1Claude Opus 5.5Max61.4%
- 2Claude Fable 5.1Max59.1%
- 3Gemini 4 ArgonHigh57.1%
- 25GLM-5.3-FlashMax39.9%
- 26Ling-3.1-flashReasoning39.4%
- 27DeepSeek-V4.1-FlashMax39.2%
- 28MiniMax-M3Reasoning39.0%
- 29GPT-6 LunaMax38.5%
- 30Agnes 3.0 FlashReasoning38.5%
- 31Qwen3.8-Flash-NextReasoning38.0%
Peers use their best measured result on this benchmark.
Alternatives
- Scores higherGemini 3.1 Pro PreviewReasoning
- Intelligence Index
- 29.7+0.5
- Cost / task
- $1.30+156%
- Higher score, lower costMiMo-V2.6-ProReasoning
- Intelligence Index
- 46.3+17.1
- Cost / task
- $0.133−74%
- Similar scoreBeam 1.0Reasoning
- Intelligence Index
- 30.3+1.1
- Cost / task
- $0.566+11%
Each model at its highest-scoring measured setting · Intelligence Index v4.3.2 and its evaluation cost per task.