0 vs 2 benchmarks won
Anthropic Claude 3.5 Haiku | Moonshot AI Kimi K2 (0905) | |
|---|---|---|
| Overview | ||
| Company | Anthropic | Moonshot AI |
| Release date | Oct 22 2024 | Sep 5 2025 |
| Model type | — | — |
| Open source | No | Yes |
| Specifications | ||
Parameters | — | 1T |
Context window | — | 256k |
| Benchmarks | ||
Science reasoning GPQA Diamond | 41.6% | 75.1% |
Software engineering SWE-Bench Verified | 40.6% | 65.8% |
Multimodal understanding MMMU | — | — |
| Timeline | ||
| Release gap | Claude 3.5 Haiku shipped 318 days before Kimi K2 (0905) | |
Kimi K2 (0905) leads Claude 3.5 Haiku on 2 of the tracked benchmarks (GPQA Diamond, SWE-Bench Verified, MMMU). Claude 3.5 Haiku shipped 318 days before Kimi K2 (0905), so benchmark comparisons should account for the intervening progress.
Kimi K2 (0905) is an open-source / open-weight model; Claude 3.5 Haiku is proprietary.
On GPQA Diamond, Kimi K2 (0905) scores 75.1%, 33.5 points above Claude 3.5 Haiku at 41.6%. On SWE-Bench Verified, Kimi K2 (0905) scores 65.8%, 25.2 points above Claude 3.5 Haiku at 40.6%.