Kimi K2.6vsMiniMax-M2.7
Across 8 shared benchmarks, Kimi K2.6 leads overall: Kimi K2.6 wins 7, MiniMax-M2.7 wins 1, with 0 ties and an average score difference of +4.22.
Kimi K2.6
Moonshot AI · 2026-04-20 · Reasoning model
MiniMax-M2.7
MiniMaxAI · 2026-03-18 · Reasoning model
Kimi K2.67 wins(88%)(13%)1 winMiniMax-M2.7
Benchmark scores
Grouped by capability, sorted by largest gap within each. 8 shared benchmarks.
Coding and Software Engineer
Kimi K2.6 3/3| Benchmark | Kimi K2.6 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| SWE-Bench Pro - Public | 58.6018 / 62Thinking (With Tools) | 56.2029 / 62Thinking (With Tools) | +2.40 |
| SciCode | 51.5061 / 130Thinking (No Tools) | 50.1068 / 130Thinking (No Tools) | +1.40 |
| SWE-bench Multilingual | 76.709 / 30Thinking (With Tools) | 76.5010 / 30Thinking (With Tools) | +0.20 |
AI Agent - Tool Usage
Kimi K2.6 2/2| Benchmark | Kimi K2.6 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| Terminal-Bench 2.1 | 65.9094 / 192Thinking (With Tools) | 55.40121 / 192Thinking (With Tools) | +10.50 |
| Terminal Bench 2.0 | 66.7010 / 48Thinking (With Tools) | 5725 / 48Thinking (With Tools) | +9.70 |
Agent Level Benchmark
Kimi K2.6 1/1| Benchmark | Kimi K2.6 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| τ³-Banking | 23.3082 / 164Thinking (With Tools) | 9.90129 / 164Thinking (With Tools) | +13.40 |
Claw-style Agent Evaluation
MiniMax-M2.7 1/1| Benchmark | Kimi K2.6 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| Claw Bench | 80.9019 / 29Thinking (With Tools) | 91.705 / 29Thinking (With Tools) | -10.80 |
Multimodal Understanding
Kimi K2.6 1/1| Benchmark | Kimi K2.6 | MiniMax-M2.7 | Diff |
|---|---|---|---|
| GDP.pdf | 1365 / 118Thinking (No Tools) | 691 / 118Thinking (No Tools) | +7 |
Specs
| Field | Kimi K2.6 | MiniMax-M2.7 |
|---|---|---|
| Publisher | Moonshot AI | MiniMaxAI |
| Release date | 2026-04-20 | 2026-03-18 |
| Model type | Reasoning model | Reasoning model |
| Architecture | MoE | MoE |
| Parameters | 1T | 229B |
| Context length | 256K | 200K |
| Max output | Not available | 200K |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | Kimi K2.6 | MiniMax-M2.7 |
|---|---|---|
| Text input | $0.95 / 1M tokens | $0.3 / 1M tokens |
| Text output | $4 / 1M tokens | $1.2 / 1M tokens |
| Cache read | $0.16 / 1M tokens | $0.06 / 1M tokens |
| Cache write | $0.95 / 1M tokens | $0.375 / 1M tokens |
Summary
- Kimi K2.6leads in:Coding and Software Engineer (3/3), AI Agent - Tool Usage (2/2), Agent Level Benchmark (1/1), Multimodal Understanding (1/1)
- MiniMax-M2.7leads in:Claw-style Agent Evaluation (1/1)
On average across the 8 shared benchmarks, Kimi K2.6 scores 4.22 higher.
Largest single-benchmark gap: τ³-Banking — Kimi K2.6 23.30 vs MiniMax-M2.7 9.90 (+13.40).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.