Claude Opus 4.7 vs GPT-6 Astra
Claude Opus 4.7 
GPT-6 Astra Verdict
- GPT-6 Astra (max) leads on quality: 70.6 vs 65.5.
- Claude Opus 4.7 is stronger in agents & tools, composite, human preference, instruction following, long context, multimodal.
- GPT-6 Astra (max) is stronger in coding, knowledge, maths, reasoning.
- Claude Opus 4.7 is 2.2× cheaper ($10.00 vs $22.00 per 1M blended).
| Metric | Claude Opus 4.7 | GPT-6 Astra (max) |
|---|---|---|
| BenchLeader Index | 65.5 | 70.6 |
| Agents & tools score | 69.4 | 68.2 |
| Coding score | 64.8 | 79.0 |
| Composite score | 80.3 | 73.1 |
| Human preference score | 68.8 | – |
| Instruction following score | 58.5 | – |
| Knowledge score | 66.4 | 79.8 |
| Long context score | 65.9 | – |
| Maths score | 65.2 | 77.0 |
| Multimodal score | 65.7 | – |
| Reasoning score | 66.3 | 70.9 |
| Blended price $/M | $10.00 | $22.00 |
| Output speed | 46 tok/s | 54 tok/s |
| Time to first answer | 21.2 s | 328.6 s |
| Context window | 1M | 1.1M |
| GPQA Diamond | – | 95.8% |
| FrontierMath Tiers 1–3 | – | 93.7% |
| FrontierMath Tier 4 | – | 97.6% |
| OTIS Mock AIME | – | 100.0% |
| SimpleQA Verified | – | 75.6% |
| Humanity's Last Exam | 36.2% | – |
| Terminal-Bench | 80.2% | 58.2% |
| SimpleBench | 61.7% | – |
| SciCode | – | 56.5% |
| WeirdML | 76.4% | – |
| FrontierCode | 38.5% | 53.3% |
| GSO-Bench | 44.1% | – |
| Epoch Capabilities Index | 156.3 | – |
| LMArena Text | 1495 | – |
| LMArena Hard Prompts | 1519 | – |
| LMArena Coding | 1547 | – |
| LMArena WebDev | 1557 | 1796 |
| LMArena Vision | 1317 | – |
| LMArena Agent | – | 12.5 |
| LiveBench | – | 82.2% |
| LiveBench Reasoning | – | 92.7% |
| LiveBench Coding | – | 80.4% |
| LiveBench Agentic Coding | – | 57.3% |
| LiveBench Mathematics | – | 96.8% |
| LiveBench Data Analysis | – | 83.0% |
| LiveBench Language | – | 89.4% |
| AA Intelligence Index | 40.7 | – |
| IFBench | 58.6% | – |
| AA-LCR | 78.7% | – |
| MMMU-Pro | 78.8% | – |
| AA-Omniscience | 27.3 | – |
| Terminal-Bench Hard | 51.5% | – |
| GPQA Diamond (AA) | 91.4% | – |
| Humanity's Last Exam (AA) | 42.3% | – |
| τ²-Bench Telecom (AA) | 88.6% | – |
| AIME (Vals) | 96.3% | – |
| LiveCodeBench | 85.1% | – |
| MMLU-Pro | 89.9% | – |
| IOI | 47.1% | – |
| LegalBench | 85.3% | – |
| CorpFin | 66.1% | – |
| TaxEval | 75.3% | – |
| Terminal-Bench 2.1 (Vals) | 68.5% | 87.3% |
| SWE-bench (Vals) | 82.0% | – |
| GPQA Diamond (Vals) | 90.2% | – |
| Vals Index | 56.1 | 66.6 |
| HiL-Bench | 41.7% | – |
| EQ-Bench 4 | 1311 | – |
| Kagi LLM Benchmark | 73.3% | – |
| ARC-AGI-1 | – | 97.5% |
| ARC-AGI-2 | – | 95.0% |
| ARC-AGI-3 | – | 62.7% |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.