| Benchmark | GLM-5.1 | o4-mini |
|---|---|---|
| AA Agentic Index | 66 | 36.1 |
| AA Intelligence | 41 | 26.1 |
| AA-LCR | 68 | 60 |
| AA-Omniscience | 0.8 | -35.7 |
| Artificial Analysis Coding Index | 55.8 | 25.6 |
| browsecomp | 79.3 | 51.5 |
| critpt | 4.6 | 0.6 |
| frontiermath_tier_4 | 12.5 | 6.3 |
| gdpval | 49.5 | 25.4 |
| GPQA Diamond | 86.8 | 81.4 |
| HLE | 52.3 | 16.5 |
| IFBench | 76.3 | 68.7 |
| OmniScience Accuracy | 25.2 | 24.8 |
| OmniScience Non-Hallucination | 70.1 | 19.5 |
| scicode | 43.8 | 46.5 |
| simplebench | 55.1 | 38.7 |
| simpleqa_verified | 38.1 | 23.9 |
| Terminal-Bench Hard | 43.2 | 15.2 |
| τ²-Bench Telecom (AA run) | 97.7 | 55.6 |
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.