| Benchmark | GLM-5.2 Full Open Source | Opus 5 |
|---|---|---|
| AA Agentic Index | 45.7 | 59.2 |
| AA Intelligence | 52.6 | 63.1 |
| AA-LCR | 76.7 | 78.7 |
| AA-Omniscience | 4.4 | 37.1 |
| arc_agi_1 | 77 | 97.5 |
| ARC-AGI-2 | 22.8 | 90.4 |
| arena_elo | 1471 | 1487 |
| arena_text_factuality | 1460 | 1481 |
| Artificial Analysis Coding Index | 68.8 | 78 |
| coding_arena_elo | 1582 | 1662 |
| critpt | 20.9 | 29.1 |
| DeepSWE 1.1 | 44 | 68.8 |
| gdpval | 50.3 | 67.2 |
| GDPval-AA v2 | 1514 | 1861 |
| GPQA Diamond | 91.2 | 93.7 |
| HiL-Bench | 43.7 | 57 |
| HLE | 54.7 | 64.7 |
| HLE (with tools) | 54.7 | 64.7 |
| livebench_agentic_coding | 51.8 | 65.2 |
| livebench_coding | 79.7 | 81.5 |
| livebench_data_analysis | 73.7 | 74.5 |
| livebench_instruction_following | 62.3 | 63.8 |
| livebench_language | 76.2 | 88.7 |
| livebench_math | 89.8 | 95.7 |
| livebench_reasoning | 78.6 | 91.2 |
| MCP Atlas | 82.6 | 85.8 |
| officeqa_pro | 41.4 | 66.9 |
| OmniScience Accuracy | 24.3 | 60.9 |
| OmniScience Non-Hallucination | 73.7 | 40.5 |
| scicode | 50.5 | 55.7 |
| simplebench | 58.8 | 80.6 |
| simpleqa_verified | 38.1 | 56.7 |
| SWE-bench Pro | 62.1 | 79.2 |
| TauBench V3 - Banking | 34.6 | 44.7 |
| Terminal-Bench 2.1 | 82.7 | 89.1 |
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.