| Benchmark | GLM-5.1 | Kimi K2 (Reasoning) |
|---|---|---|
| AA Agentic Index | 66 | 47.9 |
| AA Intelligence | 41 | 33.5 |
| AA-LCR | 68 | 70.3 |
| AA-Omniscience | 0.8 | -21.4 |
| Artificial Analysis Coding Index | 55.8 | 34.8 |
| browsecomp | 79.3 | 60.2 |
| browsecomp_with_context_manager | 79.3 | 60.2 |
| critpt | 4.6 | 2.6 |
| frontiermath_tier_4 | 12.5 | 0 |
| gdpval | 49.5 | 24.5 |
| GPQA Diamond | 86.8 | 84.5 |
| HLE | 52.3 | 23.9 |
| HLE (with tools) | 52.3 | 44.9 |
| HMMT 2025 | 94 | 89.4 |
| hmmt_nov_2025 | 94 | 89.2 |
| IFBench | 76.3 | 68.1 |
| imo_answer_bench | 83.8 | 78.6 |
| MMLU-Pro | 86 | 84.6 |
| OmniScience Accuracy | 25.2 | 30.9 |
| OmniScience Non-Hallucination | 70.1 | 25.8 |
| scicode | 43.8 | 44.8 |
| simplebench | 55.1 | 39.6 |
| SWE-bench Multilingual | 73.3 | 61.1 |
| Terminal-Bench 2.0 | 69 | 35.7 |
| Terminal-Bench Hard | 43.2 | 31.1 |
| τ²-Bench Telecom (AA run) | 97.7 | 93 |
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.