| Benchmark | Kimi K2 (Reasoning) | Muse Spark |
|---|---|---|
| AA Agentic Index | 47.9 | 28.7 |
| AA Intelligence | 33.5 | 44.3 |
| AA-LCR | 70.3 | 77 |
| AA-Omniscience | -21.4 | 7.2 |
| Artificial Analysis Coding Index | 34.8 | 58.6 |
| critpt | 2.6 | 11.3 |
| frontiermath_tier_4 | 0 | 14.6 |
| gdpval | 24.5 | 32.3 |
| GPQA Diamond | 84.5 | 89.5 |
| HLE | 23.9 | 58.4 |
| IFBench | 68.1 | 75.9 |
| multichallenge | 66.4 | 75.5 |
| OmniScience Accuracy | 30.9 | 49.6 |
| OmniScience Non-Hallucination | 25.8 | 26.8 |
| scicode | 44.8 | 51.5 |
| SWE-bench Verified | 71.3 | 77.4 |
| Terminal-Bench 2.0 | 35.7 | 59 |
| Terminal-Bench Hard | 31.1 | 45.5 |
| τ²-Bench Telecom (AA run) | 93 | 91.5 |
Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.