Compare / Claude Opus 4.5 vs o3
Claude Opus 4.5 vs o3
19 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Claude Opus 4.5 | o3 |
|---|---|---|---|
| Arena (formerly LMArena)reported | OverallArena rating | 1,187 | 1,192 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,459 | 1,405 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,472 | 1,426 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,469 | 1,383 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,503 | 1,449 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,483 | 1,403 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,474 | 1,432 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,467 | 1,396 |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Irrelevance detection% correct | 90.8% | 86.1% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Memory% correct | 73.8% | 51.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Multi-turn tasks% correct | 68.4% | 62.2% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Overall accuracy% correct | 77.5% | 63.0% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Relevance detection% correct | 68.8% | 93.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (curated)% correct | 89.7% | 81.9% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (user-contributed)% correct | 79.8% | 73.2% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Web search% correct | 84.5% | 77.0% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 3.0% | 18.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.39 | 0.30 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 70.4 | 66.8 |
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.