MiniMax M2.7
MiniMax M2.7MiniMaxOpen weightsNon-reasoningVerifiedReleased Mar 18, 2026Input: text
Overall score
48.2
Rank 111 of 222 · 36.8–59.5
Context window
200K
200,000 tokens
Price per 1M tokens
$0.30 in · $1.20 out
Source: BenchLM
Blended price
$0.53
3:1 input:output tokens
Output speed
—
Not measured yet
Category scores
Scores per capability area with the model's rank among ranked models in that category.
- 1Instruction following#1393.0
- 2Reasoning78.8
- 3Math70.5
- 4Agentic & tool use50.1
- 5Coding#8146.1
- 6Knowledge43.6
Benchmark results (22)
As reported in system cards, papers and public evaluations. Position compares this model with every other model reporting the same benchmark.
Agentic & tool use
| Gert Labs | 40.40 | #36 of 52 |
| Terminal-Bench 2.0 | 57.0 | #22 of 39 |
| Claw-Eval | 48.7 | #33 of 36 |
| τ³-bench results | 67.6 | #9 of 23 |
| Toolathlon | 46.3 | #14 of 22 |
| MLE-Bench Lite | 66.6 | #2 of 2 |
| MM-ClawBench | 62.7 | #1 of 2 |
Coding
| SWE-bench Pro | 56.2 | #43 of 69 |
| NL2Repo | 39.8 | #24 of 31 |
| React Native Evals | 71.4 | #14 of 15 |
| SWE-Rebench | 51.9 | #12 of 13 |
| SWE-bench Verified* | 75.4 | #2 of 5 |
| Multi-SWE Bench | 52.7 | #1 of 1 |
| VIBE-Pro | 55.6 | #1 of 1 |
Math
| AIME26 | 89.8 | #22 of 29 |
| HMMT Feb 2026 | 72.7 | #20 of 24 |
| HMMT Nov 2025 | 81.0 | #11 of 12 |
| AIME25 (Arcee) | 80.0 | #5 of 5 |
Knowledge
| HLE | 28.0 | #48 of 69 |
| MMLU-Pro (Arcee) | 80.8 | #5 of 5 |
Multilingual
| SWE Multilingual | 76.5 | #22 of 46 |
Multimodal
| Design Arena Website | 1243 | #48 of 91 |
Pricing by source
USD per 1M tokens.
| Source | Input | Output | Cached in |
|---|---|---|---|
| List price (BenchLM) | $0.30 | $1.20 | — |
| OpenRouter minimax/minimax-m2.7 | $0.30 | $1.20 | $0.060 |
Official MiniMax pay-as-you-go API rate for the open-weight MiniMax-M2.7 release ($0.3/$1.2 per 1M tokens; prompt caching read $0.06, write $0.375).
Arena ratings
Human-preference ratings from LMArena. Best-rated variant shown.
| Text | 1405 | #148 · 85K votes |
| WebDev (coding) | 1399 | #98 · 16K votes |
Open weights
From the Hugging Face repository.
- Parameters
- 229B
- Licence
- other
- Downloads (30 days)
- 527.8K
- Repository
- MiniMaxAI/MiniMax-M2.7 ↗
Speed
Measured by us through OpenRouter (median of trials).
Not measured yet.
Related models
Similar score
- GPT-5 mini48.4 · $0.69
- Claude Sonnet 4.548.0 · $6.00
- Qwen3.6-27B48.6 · $0.69
- Quasar 438B48.6 · $0.90
- Qwen3.5-27B48.6 · $0.74