MiniMax-M3
m3
MiniMax multimodal model for long-context coding, perception, and agent planning
Specification
most-agreed values
Context
1M
Max output
512K
Released
2026-06-01
Knowledge cutoff
—
Retires
—
Open weights
yes
Input
text, image, video
Output
text
Price
US dollars per million tokens · most-agreed
Input
$0.60
Output
$2.40
Cache read
$0.12
Quality
5 benchmarks · 2 sources
Composite
53rd
percentile of 334 scored models
Rank
159 / 334
±0.12 sd
Evidence
5 × 2
3 or more benchmarks, from 2 or more sources
Effort range
—
one configuration only
reasoning62nd3/4 bench
math79th1/7 bench
general9th1/1 bench
Every score
one row per source and configuration — nothing averaged away
| Benchmark | Score | Configuration | Source | Run |
|---|---|---|---|---|
| GPQA diamond | 90.9 ±2.0 | default | Epoch AI | 2026-08-10 |
| OTIS Mock AIME 2024-2025 | 71.1 ±6.8 | default | Epoch AI | 2026-08-10 |
| Chess Puzzles | 14.0 ±3.5 | default | Epoch AI | 2026-08-10 |
| Mystery Game Puzzles | 7.0 ±2.6 | default | Epoch AI | 2026-08-27 |
| LiveBench | 67.5 | default · 23 runs | LiveBench | 2026-06-25 |
The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.
Available from 1 host
| Host | In $/M | Out $/M | Cache rd | Cache wr | Context | Output | Retires |
|---|---|---|---|---|---|---|---|
| opper · minimax | $0.60 | $2.40 | $0.12 | — | 1M | 512K | — |