Claude Opus 4.8 vs GPT-5.6 Sol
Claude Opus 4.8 
GPT-5.6 Sol Verdict
- GPT-5.6 Sol (high) leads on quality: 68.0 vs 64.7.
- Claude Opus 4.8 is stronger in human preference, reasoning.
- GPT-5.6 Sol (high) is stronger in agents & tools, coding, composite, instruction following, knowledge, long context, multimodal.
- GPT-5.6 Sol (high) is 1.3× cheaper ($8.00 vs $10.00 per 1M blended).
| Metric | Claude Opus 4.8 | GPT-5.6 Sol (high) |
|---|---|---|
| BenchLeader Index | 64.7 | 68.0 |
| Agents & tools score | 65.0 | 87.6 |
| Coding score | 65.7 | 72.0 |
| Composite score | 81.9 | 82.5 |
| Human preference score | 65.6 | – |
| Instruction following score | 61.6 | 67.6 |
| Knowledge score | 66.3 | 73.7 |
| Long context score | 65.3 | 67.4 |
| Multimodal score | 64.4 | 66.5 |
| Reasoning score | 64.2 | 62.9 |
| Blended price $/M | $10.00 | $8.00 |
| Output speed | 58 tok/s | 59 tok/s |
| Time to first answer | 29.5 s | 31.4 s |
| Context window | 1M | 1M |
| SimpleBench | 64.8% | – |
| SciCode | – | 56.9% |
| Remote Labor Index | 8.3% | – |
| WeirdML | – | 88.8% |
| FrontierCode | 46.5% | – |
| GSO-Bench | 47.1% | – |
| Epoch Capabilities Index | 158.3 | – |
| LMArena Text | 1473 | – |
| LMArena Hard Prompts | 1503 | – |
| LMArena Coding | 1527 | – |
| LMArena WebDev | 1540 | – |
| LMArena Vision | 1289 | – |
| AA Intelligence Index | 42.0 | 42.5 |
| IFBench | 62.2% | 69.2% |
| AA-LCR | 77.7% | 81.7% |
| MMMU-Pro | – | 81.8% |
| AA-Omniscience | 28.8 | 20.4 |
| Terminal-Bench Hard | 58.3% | 62.1% |
| GPQA Diamond (AA) | 92.0% | 92.8% |
| Humanity's Last Exam (AA) | 48.7% | 46.0% |
| SciCode (AA) | 54.4% | 57.8% |
| τ²-Bench Telecom (AA) | 94.4% | 83.3% |
| LiveCodeBench | 87.8% | – |
| MMLU-Pro | 89.6% | – |
| LegalBench | 83.6% | – |
| CorpFin | 66.7% | – |
| TaxEval | 75.6% | – |
| Terminal-Bench 2.1 (Vals) | 71.9% | – |
| SWE-bench (Vals) | 88.6% | – |
| GPQA Diamond (Vals) | 92.4% | – |
| Vals Index | 60.9 | – |
| EnigmaEval | – | 37.1% |
| HiL-Bench | 35.3% | – |
| EQ-Bench 4 | 1281 | – |
| ARC-AGI-1 | – | 97.0% |
| ARC-AGI-2 | – | 85.4% |
| ARC-AGI-3 | – | 2.2% |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.