0 vs 2 benchmarks won
Anthropic Claude 3.5 Sonnet | Moonshot AI Kimi K2.5 | |
|---|---|---|
| Overview | ||
| Company | Anthropic | Moonshot AI |
| Release date | Jun 20 2024 | Jan 27 2026 |
| Model type | — | — |
| Open source | No | Yes |
| Specifications | ||
Parameters | — | 1T |
Context window | — | 256k |
| Benchmarks | ||
Science reasoning GPQA Diamond | 59.4% | 87.6% |
Software engineering SWE-Bench Verified | 33.4% | 76.8% |
Multimodal understanding MMMU | — | — |
| Timeline | ||
| Release gap | Claude 3.5 Sonnet shipped 586 days before Kimi K2.5 | |
Kimi K2.5 leads Claude 3.5 Sonnet on 2 of the tracked benchmarks (GPQA Diamond, SWE-Bench Verified, MMMU). Claude 3.5 Sonnet shipped 586 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 3.5 Sonnet is proprietary.
On GPQA Diamond, Kimi K2.5 scores 87.6%, 28.2 points above Claude 3.5 Sonnet at 59.4%. On SWE-Bench Verified, Kimi K2.5 scores 76.8%, 43.4 points above Claude 3.5 Sonnet at 33.4%.