0 vs 2 benchmarks won
Anthropic Claude Opus 4.1 | Moonshot AI Kimi K2.5 | |
|---|---|---|
| Overview | ||
| Company | Anthropic | Moonshot AI |
| Release date | Aug 5 2025 | Jan 27 2026 |
| Model type | — | — |
| Open source | No | Yes |
| Specifications | ||
Parameters | — | 1T |
Context window | — | 256k |
| Benchmarks | ||
Science reasoning GPQA Diamond | 80.9% | 87.6% |
Software engineering SWE-Bench Verified | 74.5% | 76.8% |
Multimodal understanding MMMU | — | — |
| Timeline | ||
| Release gap | Claude Opus 4.1 shipped 175 days before Kimi K2.5 | |
Kimi K2.5 leads Claude Opus 4.1 on 2 of the tracked benchmarks (GPQA Diamond, SWE-Bench Verified, MMMU). Claude Opus 4.1 shipped 175 days before Kimi K2.5, so benchmark comparisons should account for the intervening progress.
Kimi K2.5 is an open-source / open-weight model; Claude Opus 4.1 is proprietary.
On GPQA Diamond, Kimi K2.5 scores 87.6%, 6.7 points above Claude Opus 4.1 at 80.9%. On SWE-Bench Verified, Kimi K2.5 scores 76.8%, 2.3 points above Claude Opus 4.1 at 74.5%.