Claude Opus 4.6 vs Claude Opus 4.8
Claude Opus 4.6 Verdict
- Claude Opus 4.8 leads on quality: 64.7 vs 63.1.
- Claude Opus 4.6 is stronger in agents & tools, knowledge, long context, maths.
- Claude Opus 4.8 is stronger in coding, composite, human preference, instruction following, multimodal, reasoning.
- They cost about the same ($10.00 per 1M blended).
- Claude Opus 4.8 streams 1.5× faster (58 vs 38 tokens per second).
| Metric | Claude Opus 4.6 | Claude Opus 4.8 |
|---|---|---|
| BenchLeader Index | 63.1 | 64.7 |
| Agents & tools score | 68.8 | 65.0 |
| Coding score | 62.3 | 65.7 |
| Composite score | 69.2 | 81.9 |
| Human preference score | 64.6 | 65.6 |
| Instruction following score | 53.7 | 61.6 |
| Knowledge score | 70.5 | 66.3 |
| Long context score | 65.5 | 65.3 |
| Maths score | 65.6 | – |
| Multimodal score | 63.6 | 64.4 |
| Reasoning score | 62.3 | 64.2 |
| Blended price $/M | $10.00 | $10.00 |
| Output speed | 38 tok/s | 58 tok/s |
| Time to first answer | 2.0 s | 29.5 s |
| Context window | 1M | 1M |
| GPQA Diamond | 90.5% | – |
| OTIS Mock AIME | 94.4% | – |
| SWE-bench Verified (Epoch) | 78.7% | – |
| Humanity's Last Exam | 19.0% | – |
| Terminal-Bench | 79.8% | – |
| SimpleBench | 67.6% | 64.8% |
| Cybench | 93.0% | – |
| Remote Labor Index | 4.2% | 8.3% |
| WeirdML | 77.9% | – |
| APEX-Agents | 32.4% | – |
| FrontierCode | 26.6% | 46.5% |
| GSO-Bench | 33.3% | 47.1% |
| Epoch Capabilities Index | 155.3 | 158.3 |
| LMArena Text | 1498 | 1473 |
| LMArena Hard Prompts | 1527 | 1503 |
| LMArena Coding | 1546 | 1527 |
| LMArena WebDev | 1537 | 1540 |
| LMArena Vision | 1311 | 1289 |
| AA Intelligence Index | 31.9 | 42.0 |
| IFBench | 53.1% | 62.2% |
| AA-LCR | 78.0% | 77.7% |
| MMMU-Pro | 75.4% | – |
| AA-Omniscience | 13.7 | 28.8 |
| Terminal-Bench Hard | 48.5% | 58.3% |
| GPQA Diamond (AA) | 89.6% | 92.0% |
| Humanity's Last Exam (AA) | 39.9% | 48.7% |
| SciCode (AA) | – | 54.4% |
| τ²-Bench Telecom (AA) | 92.1% | 94.4% |
| 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 |
| HiL-Bench | 38.3% | 35.3% |
| EQ-Bench 4 | 1223 | 1281 |
| Kagi LLM Benchmark | 72.4% | – |
| SWE-bench Verified (bash only) | 75.6% | – |
| SWE-bench Verified (any scaffold) | 75.6% | – |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.