Model
Kimi K2 Thinking Turbo
Moonshot AI · open weights
Knowledge & Reasoning
Knowledge & ReasoningCoding
Coding| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| Aider Polyglot | 59.1 | percent_correct | N/A | Independent | medium | aider.chat |
| ARCS | 19.3 | ex_e2e | N/A | Independent | medium | arxiv.org |
| GSO | 4.9 | score_opt@1 | OpenHands | Independent | medium | gso-bench.github.io |
| WebUIProof | 39.3 | success_rate | N/A | Independent | medium | arxiv.org |
Agentic / Tool Use
Agentic / Tool Use| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| Terminal-Bench | 35.7 | accuracy | Terminus 2 | Independent | medium | tbench.ai |
| Terminal-Bench | 26.7 | accuracy | OpenHands | Independent | medium | tbench.ai |
Reproduced by the community
Loading community runs…