evals.report
BenchmarksLabsCompareRun guidesIn the wild
AnthropicClaude Opus

Claude Opus 4.1

Anthropic · Claude Opus. Released Aug 5, 2025.

Claude Opus 4.1 is a model from Anthropic in the Claude Opus family, released Aug 5, 2025. evals.report tracks 26 reported Claude Opus 4.1 benchmark scores across SWE-bench Verified, GPQA Diamond, FrontierMath, AIME (OTIS Mock), SimpleQA Verified, WeirdML, MCP Atlas, Artificial Analysis Intelligence Index, and 18 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.

26 results

Benchmark results 26

Compare this model
BenchmarkCategoryScoreMetricStatusDate
SWE-bench VerifiedCoding73.3%% resolvedOfficialAug 5, 2025Details
GPQA DiamondReasoning77.3%accuracyOfficialAug 5, 2025Details
FrontierMathReasoning7.24%accuracyOfficialAug 5, 2025Details
AIME (OTIS Mock)Reasoning68.9%accuracyOfficialAug 5, 2025Details
SimpleQA VerifiedOther34.8%accuracyOfficialAug 5, 2025Details
WeirdMLCoding45.9%average accuracyOfficialAug 5, 2025Details
MCP AtlasTool use40.9%pass rateOfficialAug 5, 2025Details
Artificial Analysis Intelligence IndexReasoning42IndexUnverifiedAug 5, 2025Details
Epoch Capabilities IndexReasoning144.7IndexOfficialAug 5, 2025Details
MMLU-ProReasoning88.0%accuracyVerifiedAug 5, 2025Details
GAIA: A Benchmark for General AI AssistantsAgents68.48%accuracyUnverifiedAug 5, 2025Details
METR Task-Completion Time HorizonsAgents100.5 min50% time horizonOfficialAug 5, 2025Details
SciCodeCoding40.9%accuracyUnverifiedAug 5, 2025Details
MMMU (Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark)Multimodal77.1%accuracyVerifiedAug 5, 2025Details
MultiChallengeReasoning57.20%accuracyVerifiedAug 5, 2025Details
WebDev ArenaChat preference1386EloVerifiedAug 5, 2025Details
Search ArenaChat preference1148EloVerifiedAug 5, 2025Details
Design ArenaChat preference1219EloVerifiedAug 5, 2025Details
MASK (Model Alignment between Statements and Knowledge)Other94.20Honesty scoreVerifiedAug 5, 2025Details
MCP-UniverseTool use29.44%Overall Success RateVerifiedAug 5, 2025Details
FACTS GroundingReasoning54.8%Grounding accuracyVerifiedAug 5, 2025Details
EnigmaEvalReasoning7.18%accuracyVerifiedAug 5, 2025Details
ZeroBenchMultimodal1.0% (pass@1)accuracyVerifiedAug 5, 2025Details
FrontierMath Tier 4Reasoning4.2%accuracyOfficialAug 5, 2025Details
Vectara Hallucination LeaderboardOther11.8%Hallucination RateOfficialAug 5, 2025Details
MultiNRCReasoning38.39%accuracyOfficialAug 5, 2025Details