Compare / Claude Opus 4.7 vs Claude Opus 4.8

Claude Opus 4.7 vs Claude Opus 4.8

20 results from 4 sources that measured both models. Each row is in the source's own units; there is no overall winner.

Claude Opus 4.7
Anthropic · profile
Claude Opus 4.8
Anthropic · profile
Results compared
20
Shared sources
4
BenchmarkMetricClaude Opus 4.7Claude Opus 4.8
Arena (formerly LMArena)reportedOverallArena rating 1,4941,486
Arena (formerly LMArena)reportedOverallArena rating 1,2121,195
Arena (formerly LMArena)reportedOverallArena rating 1,4771,469
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,5081,490
Arena (formerly LMArena)reportedCreative writingArena rating 1,4881,467
Arena (formerly LMArena)reportedExpert promptsArena rating 1,5351,525
Arena (formerly LMArena)reportedInstruction followingArena rating 1,5031,491
Arena (formerly LMArena)reportedOverallArena rating 1,5021,480
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,4961,474
OpenHands IndexreportedAverage% resolved 69.7%71.9%
OpenHands IndexreportedFront end (SWE-Bench Multimodal)% resolved 48.5%50.0%
OpenHands IndexreportedGreenfield (Commit0)% resolved 56.2%62.5%
OpenHands IndexreportedInformation gathering (GAIA)% resolved 81.2%78.8%
OpenHands IndexreportedIssue resolution (SWE-Bench)% resolved 81.6%83.8%
OpenHands IndexreportedTesting (SWT-Bench)% resolved 80.8%84.3%
tau2-benchreportedConsistency (pass^4)% of tasks 24.7%22.7%
tau2-benchreportedTask success (pass^1)% of tasks 40.2%39.7%
UGI LeaderboardreportedRequested-length error% off the requested word count 1.0%6.0%
UGI LeaderboardreportedStyle adherencescore from 0 to 1 0.410.40
UGI LeaderboardreportedWriting scorescore out of 100 61.465.9

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.