| Benchmark | Kimi K2.6 | Muse Spark 1.1 |
|---|---|---|
| AA Agentic Index | 58.7 | 39.7 |
| AA Intelligence | 45.1 | 53.2 |
| AA-LCR | 76.7 | 81.3 |
| AA-Omniscience | 6.4 | 28.1 |
| arena_elo | 1461 | 1491 |
| arena_text_factuality | 1455 | 1466 |
| Artificial Analysis Coding Index | 61.8 | 71.3 |
| baby_vision | 68.5 | 76.3 |
| coding_arena_elo | 1509 | 1539 |
| critpt | 8 | 15.1 |
| deepsearchqa_f1 | 92.5 | 84.9 |
| Finance Agent v2 | 44.9 | 57.2 |
| gdpval | 41.4 | 43.7 |
| GPQA Diamond | 91.1 | 89.8 |
| HLE | 37.5 | 62.1 |
| livebench_agentic_coding | 46.9 | 58.5 |
| livebench_coding | 78.6 | 77.2 |
| livebench_data_analysis | 65.1 | 72.5 |
| livebench_instruction_following | 64.4 | 69.6 |
| livebench_language | 75.1 | 74.3 |
| livebench_math | 84.3 | 87.1 |
| livebench_reasoning | 79.4 | 87.7 |
| MCP Atlas | 68.1 | 88.1 |
| OmniScience Accuracy | 32.8 | 52 |
| OmniScience Non-Hallucination | 60.7 | 50 |
| OSWorld-Verified | 73.1 | 80.8 |
| scicode | 53.5 | 58.2 |
| SWE-bench Pro | 58.6 | 61.5 |
| TauBench V3 - Banking | 23.3 | 31.8 |
| Terminal-Bench 2.1 | 65.9 | 80 |
| toolathlon | 50 | 75.6 |
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.