1 vs 1 benchmarks won
Anthropic Claude Haiku 4.5 | Moonshot AI Kimi K2 | |
|---|---|---|
| Overview | ||
| Company | Anthropic | Moonshot AI |
| Release date | Oct 15 2025 | Jul 11 2025 |
| Model type | — | — |
| Open source | No | Yes |
| Specifications | ||
Parameters | — | 1T |
Context window | — | 128k |
| Benchmarks | ||
Coding SWE-Bench VerifiedReal coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better. | 73.3%Best | 65.8% |
Science GPQA DiamondGraduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 73% | 75.1%Best |
| Timeline | ||
| Release gap | Kimi K2 shipped 96 days before Claude Haiku 4.5 | |
Claude Haiku 4.5 and Kimi K2 are evenly matched across the 2 benchmarks they both report (SWE-Bench Verified, GPQA Diamond). Kimi K2 shipped 96 days before Claude Haiku 4.5, so benchmark comparisons should account for the intervening progress.
Kimi K2 is an open-source / open-weight model; Claude Haiku 4.5 is proprietary.
On SWE-Bench Verified, Claude Haiku 4.5 leads at 73.3% vs Kimi K2 at 65.8%. On GPQA Diamond, Kimi K2 leads at 75.1% vs Claude Haiku 4.5 at 73%.