Compare / Claude Opus 4.6 vs GPT-5.4
Claude Opus 4.6 vs GPT-5.4
32 results from 9 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Claude Opus 4.6 | GPT-5.4 |
|---|---|---|---|
| APEX-Agentsreported | Tasks passed% of tasks | 46.3% | 52.4% |
| Arena (formerly LMArena)reported | OverallArena rating | 1,495 | 1,470 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,223 | 1,195 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,494 | 1,485 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,501 | 1,482 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,501 | 1,443 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,547 | 1,515 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,513 | 1,471 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,505 | 1,475 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,500 | 1,465 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 31.9 | 39.0 |
| Artificial Analysisreported | GPQA Diamond% of questions | 89.6% | 92.0% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 39.9% | 43.7% |
| Artificial Analysisreported | IFBench% of instructions | 53.1% | 73.9% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 78.0% | 82.0% |
| Artificial Analysisreported | Terminal-Bench Hard% of tasks | 48.5% | 57.6% |
| Artificial Analysisreported | Τ²-bench telecom% of tasks | 92.1% | 87.1% |
| OpenHands Indexreported | Average% resolved | 66.7% | 64.3% |
| OpenHands Indexreported | Front end (SWE-Bench Multimodal)% resolved | 41.8% | 36.8% |
| OpenHands Indexreported | Greenfield (Commit0)% resolved | 56.2% | 56.2% |
| OpenHands Indexreported | Information gathering (GAIA)% resolved | 80.0% | 82.4% |
| OpenHands Indexreported | Issue resolution (SWE-Bench)% resolved | 76.8% | 75.6% |
| OpenHands Indexreported | Testing (SWT-Bench)% resolved | 78.8% | 70.4% |
| SimpleQA Verified (Epoch AI)reported | Correct answers% of questions | 47.0% | 45.1% |
| tau2-benchreported | Consistency (pass^4)% of tasks | 11.3% | 21.6% |
| tau2-benchreported | Task success (pass^1)% of tasks | 27.3% | 39.4% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 3.0% | 30.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.41 | 0.40 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 70.9 | 68.8 |
| Vectara Hallucination Leaderboardreported | Answer rate% of documents | 99.8% | 99.9% |
| Vectara Hallucination Leaderboardreported | Hallucination rate% of summaries | 12.2% | 7.0% |
| Vending-Bench 2reported | Money after a yearUS dollars | $8,017.59 | $6,144.18 |
Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.