MiniMax M3vsMiniMax-M2.7
Across 6 shared benchmarks, MiniMax M3 leads overall: MiniMax M3 wins 4, MiniMax-M2.7 wins 2, with 0 ties and an average score difference of -13.74.
MiniMax M3
MiniMaxAI · 2026-06-01 · Multimodal model
MiniMax-M2.7
MiniMaxAI · 2026-03-18 · Reasoning model
MiniMax M34 wins(67%)(33%)2 winsMiniMax-M2.7
Benchmark scores
Grouped by capability, sorted by largest gap within each. 6 shared benchmarks.
Coding and Software Engineer
MiniMax M3 1/1| Benchmark | MiniMax M3 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| SWE-Bench Pro - Public | 5915 / 59Thinking (With Tools) | 56.2028 / 59Thinking (With Tools) | +2.80 |
General Evaluation
MiniMax-M2.7 1/1| Benchmark | MiniMax M3 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| GPQA Diamond | 81.31132 / 270Normal (No Tools) | 8781 / 270Thinking (No Tools) | -5.69 |
General Knowledge
MiniMax M3 1/1| Benchmark | MiniMax M3 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| LiveBench | 70.0240 / 115Deep Thinking (No Tools) | 63.4956 / 115Deep Thinking (No Tools) | +6.53 |
Long Context
MiniMax M3 1/1| Benchmark | MiniMax M3 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| AA-LCR | 80.331 / 27Thinking (No Tools) | 6914 / 27Thinking (With Tools) | +11.33 |
Productivity Knowledge
MiniMax-M2.7 1/1| Benchmark | MiniMax M3 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| GDPval-AA v2 | 1,38016 / 22Thinking (With Tools) | 1,49514 / 22Thinking (With Tools) | -115.25 |
Text Embedding
MiniMax M3 1/1| Benchmark | MiniMax M3 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| Context Arena | 51.1587 / 126Thinking (No Tools) | 33.29110 / 126Thinking (No Tools) | +17.86 |
Specs
| Field | MiniMax M3 | MiniMax-M2.7 |
|---|---|---|
| Publisher | MiniMaxAI | MiniMaxAI |
| Release date | 2026-06-01 | 2026-03-18 |
| Model type | Multimodal model | Reasoning model |
| Architecture | MoE | MoE |
| Parameters | 428B | 229B |
| Context length | 1M | 200K |
| Max output | 512K | 200K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | MiniMax M3 | MiniMax-M2.7 |
|---|---|---|
| Text input | ¥2.1 / 1M tokens | $0.3 / 1M tokens |
| Text output | ¥8.4 / 1M tokens | $1.2 / 1M tokens |
| Cache read | ¥0.42 / 1M tokens | $0.06 / 1M tokens |
| Cache write | Not public | $0.375 / 1M tokens |
Summary
- MiniMax M3leads in:Coding and Software Engineer (1/1), General Knowledge (1/1), Long Context (1/1), Text Embedding (1/1)
- MiniMax-M2.7leads in:General Evaluation (1/1), Productivity Knowledge (1/1)
On average across the 6 shared benchmarks, MiniMax-M2.7 scores 13.74 higher.
Largest single-benchmark gap: GDPval-AA v2 — MiniMax M3 1,380 vs MiniMax-M2.7 1,495 (-115.25).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.