MiniMax · MiniMax M3
MiniMax's current flagship LLM: an open-weight 428B/23B MoE with MiniMax Sparse Attention, 1M context and native image/video input, at a low per-token price.
Leader: Claude Opus 5.5 at 58
Leader: Claude Mythos 5.1 at 60.9%
Leader: Claude Fable 5.1 at 91.4%
Leader: Claude Opus 5.5 at 66.9%
Leader: Claude Opus 5.5 at 89.9%
Leader: GLM-5.2 at 99.1%
Leader: Claude Opus 5.5 at 1846
Leader: Kimi K3 at 84.2%
Leader: Kimi K3 at 84.8%
Leader: GPT-6 Astra at 96.1%
Leader: Claude Opus 5.5 at 61.4%
Leader: GPT-5.6 Sol at 32.3%
Leader: Kimi K3 at 88.7%
Leader: Claude Opus 5.5 at 87.7%
Leader: Claude Fable 5 at 1506
Leader: Claude Opus 5.5 at 1818
| Result | Score | Reported by | Source |
|---|---|---|---|
| IFBench (AA) · reasoning | 82.9% | independent | Artificial Analysis: MiniMax-M3 ↗ |
| Terminal-Bench Hard (AA) · reasoning | 42.4% | independent | Artificial Analysis: MiniMax-M3 ↗ |
| SWE-fficiency | 34.8% | vendor | MiniMax blog: MiniMax M3 ↗ |
| KernelBench Hard | 28.8% | vendor | MiniMax blog: MiniMax M3 ↗ |