| Benchmark | Gemini 2.5 Flash | GLM-5.1 |
|---|---|---|
| AA Agentic Index | 18.8 | 66 |
| AA Intelligence | 17.7 | 41 |
| AA-LCR | 65.7 | 68 |
| AA-Omniscience | -29.8 | 0.8 |
| arena_text_factuality | 1437 | 1458 |
| Artificial Analysis Coding Index | 22.2 | 55.8 |
| critpt | 1.4 | 4.6 |
| frontiermath_tier_4 | 4.2 | 12.5 |
| gdpval | 11.9 | 49.5 |
| GPQA Diamond | 82.8 | 86.8 |
| HLE | 12.1 | 52.3 |
| IFBench | 50.3 | 76.3 |
| OmniScience Accuracy | 26.1 | 25.2 |
| OmniScience Non-Hallucination | 24.7 | 70.1 |
| scicode | 35.9 | 43.8 |
| simplebench | 41.2 | 55.1 |
| Terminal-Bench Hard | 13.6 | 43.2 |
| τ²-Bench Telecom (AA run) | 31.6 | 97.7 |
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.