Moonshot AIKimi
Kimi K2 Instruct
Moonshot AI · Kimi. Released Jul 11, 2025.
Kimi K2 Instruct is a model from Moonshot AI in the Kimi family, released Jul 11, 2025. evals.report tracks 19 reported Kimi K2 Instruct benchmark scores across Berkeley Function Calling Leaderboard, SWE-bench Pro, WeirdML, MCP Atlas, Artificial Analysis Intelligence Index, Epoch Capabilities Index, Aider Polyglot, MMLU-Pro, and 11 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.
Open19 results
Benchmark results 19
Compare this model| Benchmark | Category | Score | Metric | Status | Date | |
|---|---|---|---|---|---|---|
| Berkeley Function Calling Leaderboard | Tool use | 59.06% | accuracy | Official | Jul 11, 2025 | Details |
| SWE-bench Pro | Coding | 27.67% | % resolved | Official | Jul 11, 2025 | Details |
| WeirdML | Coding | 39.4% | average accuracy | Official | Jul 11, 2025 | Details |
| MCP Atlas | Tool use | 23.9% | pass rate | Official | Jul 11, 2025 | Details |
| Artificial Analysis Intelligence Index | Reasoning | 26.3 | Index | Unverified | Jul 11, 2025 | Details |
| Epoch Capabilities Index | Reasoning | 141.2 | Index | Official | Jul 11, 2025 | Details |
| Aider Polyglot | Coding | 60.0% | % correct | Unverified | Jul 11, 2025 | Details |
| MMLU-Pro | Reasoning | 84.8% | accuracy | Verified | Jul 11, 2025 | Details |
| τ²-bench (Telecom) | Tool use | 65.8% | pass^1 | Unverified | Jul 11, 2025 | Details |
| LiveCodeBench | Coding | 85.3% | Pass@1 | Unverified | Jul 11, 2025 | Details |
| SciCode | Coding | 42.4% | accuracy | Unverified | Jul 11, 2025 | Details |
| LongBench v2 | Reasoning | 44.3% | accuracy | Official | Jul 11, 2025 | Details |
| EQ-Bench Creative Writing v3 | Chat preference | 1738 | Elo | Verified | Jul 11, 2025 | Details |
| Design Arena | Chat preference | 1088 | Elo | Verified | Jul 11, 2025 | Details |
| MASK (Model Alignment between Statements and Knowledge) | Other | 46.67 | Honesty score | Verified | Jul 11, 2025 | Details |
| SuperGPQA | Reasoning | 57.2% | accuracy | Unverified | Jul 11, 2025 | Details |
| IMO-Bench | Reasoning | 45.8% | accuracy | Verified | Jul 11, 2025 | Details |
| PolyMath: Evaluating Mathematical Reasoning in Multilingual Contexts | Reasoning | 65.1% | Difficulty-Weighted Accuracy (DW-ACC) | Unverified | Jul 11, 2025 | Details |
| GSO: Software Optimization Benchmark for SWE-Agents | Coding | 4.9% | Opt@1 | Official | Jul 11, 2025 | Details |