Claude Opus 5 vs GPT-5.4
Claude Opus 5 
GPT-5.4 Verdict
- Claude Opus 5 leads on quality: 70.0 vs 63.4.
- Claude Opus 5 is stronger in agents & tools, coding, composite, human preference, knowledge, maths, multimodal, reasoning.
- GPT-5.4 is stronger in long context, instruction following.
- GPT-5.4 is 1.8× cheaper ($5.63 vs $10.00 per 1M blended).
- GPT-5.4 streams 2.5× faster (130 vs 52 tokens per second).
| Metric | Claude Opus 5 | GPT-5.4 |
|---|---|---|
| BenchLeader Index | 70.0 | 63.4 |
| Agents & tools score | 71.6 | 66.1 |
| Coding score | 74.4 | 51.6 |
| Composite score | 92.9 | 78.1 |
| Human preference score | 78.2 | 64.5 |
| Knowledge score | 70.5 | 66.8 |
| Long context score | 66.2 | 67.6 |
| Maths score | 67.1 | – |
| Multimodal score | 69.4 | 64.0 |
| Reasoning score | 71.1 | 62.3 |
| Instruction following score | – | 71.7 |
| Blended price $/M | $10.00 | $5.63 |
| Output speed | 52 tok/s | 130 tok/s |
| Time to first answer | 81.6 s | 126.4 s |
| Context window | 1M | 1.1M |
| GPQA Diamond | 92.9% | – |
| OTIS Mock AIME | 97.8% | – |
| Terminal-Bench | – | 81.8% |
| SimpleBench | 80.6% | – |
| WeirdML | 86.3% | – |
| APEX-Agents | – | 34.9% |
| Epoch Capabilities Index | 162.6 | 156.9 |
| LMArena Text | – | 1466 |
| LMArena Hard Prompts | – | 1488 |
| LMArena Coding | – | 1514 |
| LMArena WebDev | – | 1386 |
| LMArena Vision | – | 1293 |
| AA Intelligence Index | 50.7 | 39.0 |
| IFBench | – | 74.0% |
| AA-LCR | 79.3% | 82.0% |
| MMMU-Pro | 84.7% | 78.4% |
| AA-Omniscience | 37.1 | 5.8 |
| Terminal-Bench Hard | – | 57.6% |
| GPQA Diamond (AA) | 93.2% | 92.0% |
| Humanity's Last Exam (AA) | 54.9% | 43.7% |
| SciCode (AA) | 56.4% | – |
| τ²-Bench Telecom (AA) | – | 87.1% |
| LiveCodeBench | 89.0% | – |
| MMLU-Pro | 91.6% | – |
| IOI | 91.7% | – |
| LegalBench | 87.0% | – |
| CorpFin | 73.2% | – |
| TaxEval | 75.1% | – |
| Terminal-Bench 2.1 (Vals) | 84.6% | – |
| SWE-bench (Vals) | 97.0% | – |
| GPQA Diamond (Vals) | 93.4% | – |
| Vals Index | 67.2 | – |
| HiL-Bench | 57.0% | 9.7% |
| EQ-Bench 4 | 1385 | 1272 |
| Kagi LLM Benchmark | – | 63.8% |
Data as of 2026-09-09. Best configuration of each model; every score links to its source on the model pages.