Claude Opus 4.6 vs GPT-5.2
Claude Opus 4.6 
GPT-5.2 Verdict
- Claude Opus 4.6 leads on quality: 63.3 vs 61.6.
- Claude Opus 4.6 is stronger in agents & tools, coding, composite, knowledge, maths, multimodal, reasoning.
- GPT-5.2 is stronger in human preference, instruction following, long context.
- GPT-5.2 is 2.1× cheaper ($4.81 vs $10.00 per 1M blended).
- GPT-5.2 streams 1.9× faster (68 vs 36 tokens per second).
| Metric | Claude Opus 4.6 | GPT-5.2 |
|---|---|---|
| BenchLeader Index | 63.3 | 61.6 |
| Agents & tools score | 68.8 | 61.2 |
| Coding score | 62.4 | 53.7 |
| Composite score | 69.4 | 67.5 |
| Human preference score | 64.9 | 67.9 |
| Instruction following score | 53.9 | 67.4 |
| Knowledge score | 70.6 | 61.1 |
| Long context score | 65.5 | 67.9 |
| Maths score | 65.7 | – |
| Multimodal score | 63.6 | 59.9 |
| Reasoning score | 62.6 | 61.3 |
| Blended price $/M | $10.00 | $4.81 |
| Output speed | 36 tok/s | 68 tok/s |
| Time to first answer | 2.4 s | 129.4 s |
| Context window | 1M | 400k |
| GPQA Diamond | 90.5% | – |
| OTIS Mock AIME | 94.4% | – |
| SWE-bench Verified (Epoch) | 78.7% | – |
| Humanity's Last Exam | 19.0% | 27.8% |
| Terminal-Bench | 79.8% | 64.9% |
| SimpleBench | 67.6% | 45.8% |
| Cybench | 93.0% | – |
| Remote Labor Index | 4.2% | 2.1% |
| WeirdML | 77.9% | – |
| APEX-Agents | 32.4% | 23.0% |
| FrontierCode | 26.6% | – |
| GSO-Bench | 33.3% | – |
| Epoch Capabilities Index | 155.3 | 153.5 |
| LMArena Text | 1498 | 1476 |
| LMArena Hard Prompts | 1527 | 1497 |
| LMArena Coding | 1546 | 1515 |
| LMArena WebDev | 1537 | 1417 |
| LMArena Vision | 1311 | 1268 |
| AA Intelligence Index | 31.9 | 30.4 |
| IFBench | 53.1% | 75.4% |
| AA-LCR | 78.0% | 82.7% |
| MMMU-Pro | 75.4% | – |
| AA-Omniscience | 13.7 | -0.9 |
| Terminal-Bench Hard | 48.5% | 47.0% |
| GPQA Diamond (AA) | 89.6% | 90.3% |
| Humanity's Last Exam (AA) | 39.9% | 37.7% |
| τ²-Bench Telecom (AA) | 92.1% | 84.8% |
| SWE-Bench Pro | – | 29.9% |
| VISTA | – | 46.6% |
| MultiNRC | – | 42.2% |
| HiL-Bench | 38.3% | – |
| TutorBench | – | 53.5% |
| EQ-Bench 4 | 1223 | – |
| Kagi LLM Benchmark | 72.4% | 73.3% |
| SWE-bench Verified (bash only) | 75.6% | 69.0% |
| SWE-bench Verified (any scaffold) | 75.6% | – |
| ARC-AGI-1 | – | 94.5% |
| ARC-AGI-2 | – | 72.9% |
| BFCL Overall | – | 55.9% |
Data as of 2026-09-10. Best configuration of each model; every score links to its source on the model pages.
Claude Opus 4.6 vs GPT-5.2: questions
- Is Claude Opus 4.6 better than GPT-5.2?
- Claude Opus 4.6 leads on quality: 63.3 vs 61.6. The BenchLeader Index combines every independent quality benchmark; Claude Opus 4.6 is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is Claude Opus 4.6 better than GPT-5.2 for coding?
- Claude Opus 4.6 scores higher in coding (62 vs 54 on the category index, where 50 is average).
- Is Claude Opus 4.6 better than GPT-5.2 for agentic tasks?
- Claude Opus 4.6 scores higher in agentic tasks (69 vs 61 on the category index, where 50 is average).
- Which is cheaper, Claude Opus 4.6 or GPT-5.2?
- GPT-5.2 is cheaper: $4.81 against $10.00 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Opus 4.6 or GPT-5.2?
- GPT-5.2 streams faster: 68 against 36 output tokens per second.
- Which has the larger context window?
- Claude Opus 4.6 accepts more context: 1M against 400k tokens.