Compare / Claude Opus 4.6 vs o3

Claude Opus 4.6 vs o3

12 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.

Claude Opus 4.6
Anthropic · profile
o3
OpenAI · profile
Results compared
12
Shared sources
3
BenchmarkMetricClaude Opus 4.6o3
Arena (formerly LMArena)reportedOverallArena rating 1,2231,192
Arena (formerly LMArena)reportedOverallArena rating 1,4941,405
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,5011,426
Arena (formerly LMArena)reportedCreative writingArena rating 1,5011,383
Arena (formerly LMArena)reportedExpert promptsArena rating 1,5471,449
Arena (formerly LMArena)reportedInstruction followingArena rating 1,5131,403
Arena (formerly LMArena)reportedOverallArena rating 1,5051,432
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,5001,396
SimpleQA Verified (Epoch AI)reportedCorrect answers% of questions 47.0%49.4%
UGI LeaderboardreportedRequested-length error% off the requested word count 3.0%18.0%
UGI LeaderboardreportedStyle adherencescore from 0 to 1 0.410.30
UGI LeaderboardreportedWriting scorescore out of 100 70.966.8

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.