Up to four models
Claude Opus 4.8 vs GLM-5.2
On the index (GPT-4 = 100), Claude Opus 4.8 scores 170 and GLM-5.2 154. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
Claude Opus 4.8
170154–186
WeightsClosed
ReleasedMay 2026
Price in / out$5.0 / $25
Context1M
Core tests taken6 of 9
154138–169
WeightsOpen
ReleasedJun 2026
Price in / out$0.0070 / $6.8
Context1.05M
Core tests taken6 of 9
Claude Opus 4.8 leads GLM-5.2 by 17 points, ahead on 4 of 5 shared core tests, at 5.9× the price.
Test by testthird-party results
Claude Opus 4.8
GLM-5.2
Every result as a table
| Test | Claude Opus 4.8 | GLM-5.2 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 91.0% | 91.9% |
| LiveBench Data Analysis | 66.0% | 73.7% |
| LiveBench IF | 72.0% | 62.3% |
| LiveBench Language | 79.7% | 76.2% |
| LiveBench Reasoning | 89.2% | 78.6% |
| GPQA Diamond (Vals AI) | 92.4% | 85.6% |
| MMLU Pro | 89.6% | 86.7% |
| EBR-bench | 28.6% | 9.5% |
| SimpleBench | 64.8% | 58.8% |
| SimpleQA Verified | 53.0% | 34.2% |
| Maths | ||
| FrontierMath (tiers 1-3) | 80.0% | 59.2% |
| AIME-style maths (OTIS mock) | 98.3% | 86.4% |
| LiveBench Mathematics | 94.3% | 89.8% |
| Chess puzzles | 52.0% | 22.0% |
| FrontierMath (tier 4) | 56.1% | 29.3% |
| Surface Evolver Bench | 87.5% | 55.6% |
| Coding | ||
| LiveBench Agentic Coding | 50.5% | 51.8% |
| LiveBench Coding | 81.8% | 79.7% |
| Code Migration | 47.3% | 37.9% |
| LiveCodeBench | 87.8% | 69.5% |
| ProgramBench | 1.0% | 0.5% |
| SWE-bench | 88.6% | 82.8% |
| Vibe Code Bench v1.1 | 82.7% | 64.0% |
| DeepSWE | 59.0% | 43.8% |
| MirrorCode | 65.6% | 36.7% |
| WeirdML | 82.9% | 70.1% |
| Agents | ||
| APEX-Agents | 48.9% | 45.2% |
| SkillsBench | 59.2% | 45.1% |
| DTBench | 94.9% | 93.6% |
| LMCA | 57.5% | 45.8% |
| Mystery game puzzles | 36.0% | 19.0% |
| Professional work | ||
| Finance Agent (v2) | 53.9% | 49.7% |
| CorpFin v2 | 66.7% | 66.1% |
| EMB | 69.4% | 61.5% |
| LegalBench | 83.6% | 84.1% |
| Legal Research Bench | 43.8% | 31.3% |
| MedCode | 53.2% | 40.8% |
| MedScribe | 85.8% | 83.5% |
| TaxEval v2 | 75.6% | 73.3% |
| Novel problems | ||
| ARC-AGI-2 | 72.1% | 22.8% |
| ARC-AGI-1 | 92.5% | 77.0% |
Tests at least two of these models have taken. Bold rows are the core tests.