| Benchmark | Kimi K3 | Qwen3.6 Plus |
|---|---|---|
| AA Agentic Index | 54.3 | 29 |
| AA Intelligence | 60 | 40.5 |
| AA-LCR | 82.7 | 72.3 |
| AA-Omniscience | 46 | 2.6 |
| aime_2026 | 96.7 | 95.3 |
| arena_elo | 1489 | 1444 |
| arena_text_factuality | 1471 | 1438 |
| Artificial Analysis Coding Index | 76.2 | 54.5 |
| coding_arena_elo | 1674 | 1460 |
| critpt | 23.4 | 2.9 |
| Finance Agent v2 | 54.4 | 40.9 |
| gdpval | 59.4 | 32 |
| GPQA Diamond | 93.5 | 90.4 |
| HLE | 56 | 28.8 |
| hmmt_feb_2026 | 97 | 87.8 |
| livebench_agentic_coding | 62.2 | 41.4 |
| livebench_coding | 81.5 | 78.2 |
| livebench_data_analysis | 78.7 | 69.9 |
| livebench_instruction_following | 71.4 | 58.3 |
| livebench_language | 85.5 | 75 |
| livebench_math | 84.4 | 83.7 |
| livebench_reasoning | 90.7 | 75.8 |
| mathvision | 97.8 | 88 |
| MCP Atlas | 84.2 | 74.1 |
| MMMU-Pro | 83.4 | 78.8 |
| OmniScience Accuracy | 47.6 | 26.4 |
| OmniScience Non-Hallucination | 49.1 | 68 |
| OSWorld-Verified | 84.8 | 62.5 |
| scicode | 58.7 | 40.7 |
| simpleqa_verified | 42.7 | 49.1 |
| TauBench V3 - Banking | 46 | 20.8 |
| Terminal-Bench 2.1 | 88.3 | 61.4 |
| toolathlon | 73.2 | 39.8 |
| Video-MME | 90 | 84.2 |
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.