MiniMax M3vsKimi K2.6
Across 6 shared benchmarks, MiniMax M3 leads overall: MiniMax M3 wins 4, Kimi K2.6 wins 2, with 0 ties and an average score difference of +2.11.
MiniMax M3
MiniMaxAI · 2026-06-01 · Multimodal model
Kimi K2.6
Moonshot AI · 2026-04-20 · Reasoning model
MiniMax M34 wins(67%)(33%)2 winsKimi K2.6
Benchmark scores
Grouped by capability, sorted by largest gap within each. 6 shared benchmarks.
AI Agent - Tool Usage
MiniMax M3 2/3| Benchmark | MiniMax M3 | Kimi K2.6 | Diff |
|---|---|---|---|
| Terminal-Bench 2.1 | 6635 / 44Thinking (With Tools) | 53.5643 / 44Thinking (No Tools) | +12.44 |
| MCP-Atlas | 74.2022 / 38Thinking (With Tools) | 69.4028 / 38Thinking (With Tools) | +4.80 |
| OSWorld-Verified | 7019 / 26Thinking (With Tools) | 73.1015 / 26Thinking (With Tools) | -3.10 |
AI Agent - Information Search
MiniMax M3 1/1| Benchmark | MiniMax M3 | Kimi K2.6 | Diff |
|---|---|---|---|
| BrowseComp | 83.5012 / 54Thinking (With Tools + Internet) | 83.2014 / 54Thinking (With Tools + Internet) | +0.30 |
Coding and Software Engineer
MiniMax M3 1/1| Benchmark | MiniMax M3 | Kimi K2.6 | Diff |
|---|---|---|---|
| SWE-Bench Pro - Public | 5913 / 57Thinking (With Tools) | 58.6015 / 57Thinking (With Tools) | +0.40 |
General Knowledge
Kimi K2.6 1/1| Benchmark | MiniMax M3 | Kimi K2.6 | Diff |
|---|---|---|---|
| LiveBench | 70.0240 / 115Deep Thinking (No Tools) | 72.1728 / 115Thinking (No Tools) | -2.15 |
Specs
| Field | MiniMax M3 | Kimi K2.6 |
|---|---|---|
| Publisher | MiniMaxAI | Moonshot AI |
| Release date | 2026-06-01 | 2026-04-20 |
| Model type | Multimodal model | Reasoning model |
| Architecture | MoE | MoE |
| Parameters | 428B | 1T |
| Context length | 1M | 256K |
| Max output | 512K | Not available |
API pricing
Prices use DataLearner records when available; missing fields are not inferred.
| Item | MiniMax M3 | Kimi K2.6 |
|---|---|---|
| Text input | ¥2.1 / 1M tokens | $0.95 / 1M tokens |
| Text output | ¥8.4 / 1M tokens | $4 / 1M tokens |
| Cache read | ¥0.42 / 1M tokens | $0.16 / 1M tokens |
| Cache write | Not public | $0.95 / 1M tokens |
Summary
- MiniMax M3leads in:AI Agent - Tool Usage (2/3), AI Agent - Information Search (1/1), Coding and Software Engineer (1/1)
- Kimi K2.6leads in:General Knowledge (1/1)
On average across the 6 shared benchmarks, MiniMax M3 scores 2.11 higher.
Largest single-benchmark gap: Terminal-Bench 2.1 — MiniMax M3 66 vs Kimi K2.6 53.56 (+12.44).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.