Claude Opus 4.8 vs GPT-5.6 Terra
Claude Opus 4.8 
GPT-5.6 Terra Verdict
- Claude Opus 4.8 and GPT-5.6 Terra (xhigh) are level on quality (64.7 vs 64.2).
- Claude Opus 4.8 is stronger in coding, composite, knowledge, multimodal, reasoning.
- GPT-5.6 Terra (xhigh) is stronger in agents & tools, human preference, instruction following, long context, maths.
- GPT-5.6 Terra (xhigh) is 2.2× cheaper ($4.50 vs $10.00 per 1M blended).
- GPT-5.6 Terra (xhigh) streams 1.6× faster (93 vs 58 tokens per second).
| Metric | Claude Opus 4.8 | GPT-5.6 Terra (xhigh) |
|---|---|---|
| BenchLeader Index | 64.7 | 64.2 |
| Agents & tools score | 65.0 | 69.6 |
| Coding score | 65.7 | 63.7 |
| Composite score | 81.9 | 77.1 |
| Human preference score | 65.6 | 66.3 |
| Instruction following score | 61.6 | 65.0 |
| Knowledge score | 66.3 | 60.9 |
| Long context score | 65.3 | 66.0 |
| Multimodal score | 64.4 | 63.0 |
| Reasoning score | 64.2 | 58.2 |
| Maths score | – | 71.0 |
| Blended price $/M | $10.00 | $4.50 |
| Output speed | 58 tok/s | 93 tok/s |
| Time to first answer | 29.5 s | 32.5 s |
| Context window | 1M | 1M |
| SimpleBench | 64.8% | 48.9% |
| SciCode | – | 51.6% |
| Remote Labor Index | 8.3% | – |
| FrontierCode | 46.5% | – |
| ProofBench | – | 74.0% |
| GSO-Bench | 47.1% | – |
| Epoch Capabilities Index | 158.3 | – |
| LMArena Text | 1473 | 1466 |
| LMArena Hard Prompts | 1503 | 1491 |
| LMArena Coding | 1527 | 1519 |
| LMArena WebDev | 1540 | 1521 |
| LMArena Vision | 1289 | 1269 |
| LMArena Agent | – | 1.5 |
| AA Intelligence Index | 42.0 | 38.2 |
| IFBench | 62.2% | 66.3% |
| AA-LCR | 77.7% | 79.0% |
| MMMU-Pro | – | 79.5% |
| AA-Omniscience | 28.8 | -3.0 |
| Terminal-Bench Hard | 58.3% | 62.9% |
| GPQA Diamond (AA) | 92.0% | 90.8% |
| Humanity's Last Exam (AA) | 48.7% | 41.9% |
| SciCode (AA) | 54.4% | 52.3% |
| τ²-Bench Telecom (AA) | 94.4% | 80.4% |
| LiveCodeBench | 87.8% | 85.9% |
| MMLU-Pro | 89.6% | 86.7% |
| IOI | – | 65.3% |
| LegalBench | 83.6% | – |
| CorpFin | 66.7% | 65.3% |
| TaxEval | 75.6% | 76.2% |
| Terminal-Bench 2.1 (Vals) | 71.9% | – |
| SWE-bench (Vals) | 88.6% | – |
| GPQA Diamond (Vals) | 92.4% | 90.9% |
| Vals Index | 60.9 | – |
| HiL-Bench | 35.3% | – |
| EQ-Bench 4 | 1281 | – |
| ARC-AGI-1 | – | 94.0% |
| ARC-AGI-2 | – | 74.2% |
| ARC-AGI-3 | – | 0.7% |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.