MiniMax-M2.7vsKimi K2.5
Across 7 shared benchmarks, MiniMax-M2.7 leads overall: MiniMax-M2.7 wins 5, Kimi K2.5 wins 2, with 0 ties and an average score difference of +1.29.
MiniMax-M2.7
MiniMaxAI · 2026-03-18 · Reasoning model
Kimi K2.5
Moonshot AI · 2026-01-27 · Multimodal model
MiniMax-M2.75 wins(71%)(29%)2 winsKimi K2.5
Benchmark scores
Grouped by capability, sorted by largest gap within each. 7 shared benchmarks.
Claw-style Agent Evaluation
MiniMax-M2.7 2/2| Benchmark | MiniMax-M2.7 | Kimi K2.5 | Diff |
|---|---|---|---|
| Claw Bench | 91.705 / 29Thinking (With Tools) | 81.7018 / 29Thinking (With Tools) | +10 |
| Pinch Bench | 87.109 / 37Thinking (With Tools) | 84.8017 / 37Thinking (With Tools) | +2.30 |
General Knowledge
Kimi K2.5 2/2| Benchmark | MiniMax-M2.7 | Kimi K2.5 | Diff |
|---|---|---|---|
| HLE | 2882 / 157Thinking (No Tools) | 50.2020 / 157Thinking (With Tools) | -22.20 |
| GPQA Diamond | 8738 / 178Thinking (No Tools) | 87.6034 / 178Thinking (No Tools) | -0.60 |
Coding and Software Engineer
MiniMax-M2.7 1/1| Benchmark | MiniMax-M2.7 | Kimi K2.5 | Diff |
|---|---|---|---|
| SWE-Bench Pro - Public | 56.2016 / 43Thinking (With Tools) | 50.7031 / 43Thinking (With Tools) | +5.50 |
Long Context
MiniMax-M2.7 1/1| Benchmark | MiniMax-M2.7 | Kimi K2.5 | Diff |
|---|---|---|---|
| AA-LCR | 694 / 13Thinking (With Tools) | 6510 / 13Thinking (No Tools) | +4 |
Productivity Knowledge
MiniMax-M2.7 1/1| Benchmark | MiniMax-M2.7 | Kimi K2.5 | Diff |
|---|---|---|---|
| GDPval-AA | 5013 / 21Thinking (No Tools) | 4015 / 21Thinking (No Tools) | +10 |
Specs
| Field | MiniMax-M2.7 | Kimi K2.5 |
|---|---|---|
| Publisher | MiniMaxAI | Moonshot AI |
| Release date | 2026-03-18 | 2026-01-27 |
| Model type | Reasoning model | Multimodal model |
| Architecture | MoE | MoE |
| Parameters | 229B | 1T |
| Context length | 200K | 256K |
| Max output | 200K | 16K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | MiniMax-M2.7 | Kimi K2.5 |
|---|---|---|
| Text input | $0.3 / 1M tokens | Not public |
| Text output | $1.2 / 1M tokens | Not public |
| Cache read | $0.06 / 1M tokens | Not public |
| Cache write | $0.375 / 1M tokens | Not public |
One or both models have incomplete public pricing.
Summary
- MiniMax-M2.7leads in:Claw-style Agent Evaluation (2/2), Coding and Software Engineer (1/1), Long Context (1/1), Productivity Knowledge (1/1)
- Kimi K2.5leads in:General Knowledge (2/2)
On average across the 7 shared benchmarks, MiniMax-M2.7 scores 1.29 higher.
Largest single-benchmark gap: HLE — MiniMax-M2.7 28 vs Kimi K2.5 50.20 (-22.20).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.