Benchmarks

21 results across 9 benchmarks

Benchmark results for Claude Sonnet 4
BenchmarkScoreSource
GPQA DiamondEffort: NoneReported as: standard (no extended thinking)70.0%anthropic.com for Claude Sonnet 4 (standard (no extended thinking)) on Graduate-Level Google-Proof Q&A Diamond (opens in a new tab)GPQA Diamond AccuracyReported Launch postAnthropicSelf-reportedRecord for Claude Sonnet 4 on Graduate-Level Google-Proof Q&A Diamond
GPQA Diamond — Epoch AI harnessConfiguration: Epoch AI harness66.7%epoch.ai for Claude Sonnet 4 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness (opens in a new tab)Archived copy (opens in a new tab)GPQA Diamond AccuracyReported Evaluated Benchmark leaderboardEpoch AIIndependentRecord for Claude Sonnet 4 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness
GPQA Diamond — Vals AI harnessConfiguration: Vals AI harness: mean of zero-shot and few-shot CoT69.4%vals.ai for Claude Sonnet 4 on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness (opens in a new tab)Archived copy (opens in a new tab)GPQA Diamond AccuracyReported Benchmark leaderboardVals AIIndependentRecord for Claude Sonnet 4 on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness
GPQA Diamond — Vals AI harnessReported as: thinkingConfiguration: Vals AI harness: mean of zero-shot and few-shot CoT75.0%vals.ai for Claude Sonnet 4 (thinking) on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness (opens in a new tab)Archived copy (opens in a new tab)GPQA Diamond AccuracyReported Benchmark leaderboardVals AIIndependentRecord for Claude Sonnet 4 on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness
Humanity's Last Exam Full set — Scale AI evaluation5.5%labs.scale.com for Claude Sonnet 4 on Humanity's Last Exam Full set — Scale AI evaluation (opens in a new tab)Archived copy (opens in a new tab)Humanity's Last Exam AccuracyReported Benchmark leaderboardScale AIIndependentRecord for Claude Sonnet 4 on Humanity's Last Exam Full set — Scale AI evaluation
Humanity's Last Exam Full set — Scale AI evaluationReported as: Thinking7.8%labs.scale.com for Claude Sonnet 4 (Thinking) on Humanity's Last Exam Full set — Scale AI evaluation (opens in a new tab)Archived copy (opens in a new tab)Humanity's Last Exam AccuracyReported Benchmark leaderboardScale AIIndependentRecord for Claude Sonnet 4 on Humanity's Last Exam Full set — Scale AI evaluation
Humanity's Last Exam Text only — Scale AI evaluation5.4%labs.scale.com for Claude Sonnet 4 on Humanity's Last Exam Text only — Scale AI evaluation (opens in a new tab)Archived copy (opens in a new tab)Humanity's Last Exam AccuracyReported Benchmark leaderboardScale AIIndependentRecord for Claude Sonnet 4 on Humanity's Last Exam Text only — Scale AI evaluation
Humanity's Last Exam Text only — Scale AI evaluationReported as: Thinking7.6%labs.scale.com for Claude Sonnet 4 (Thinking) on Humanity's Last Exam Text only — Scale AI evaluation (opens in a new tab)Archived copy (opens in a new tab)Humanity's Last Exam AccuracyReported Benchmark leaderboardScale AIIndependentRecord for Claude Sonnet 4 on Humanity's Last Exam Text only — Scale AI evaluation
MMMUEffort: NoneReported as: standard (no extended thinking)72.6%anthropic.com for Claude Sonnet 4 (standard (no extended thinking)) on Massive Multi-discipline Multimodal Understanding MMMU (opens in a new tab)MMMU AccuracyReported Launch postAnthropicSelf-reportedRecord for Claude Sonnet 4 on Massive Multi-discipline Multimodal Understanding MMMU
MMLU-Pro Original — Vals AI harnessConfiguration: Vals AI harness79.4%vals.ai for Claude Sonnet 4 on MMLU-Pro Original — Vals AI harness (opens in a new tab)Archived copy (opens in a new tab)AccuracyReported Benchmark leaderboardVals AIIndependentRecord for Claude Sonnet 4 on MMLU-Pro Original — Vals AI harness
MMLU-Pro Original — Vals AI harnessReported as: thinkingConfiguration: Vals AI harness83.9%vals.ai for Claude Sonnet 4 (thinking) on MMLU-Pro Original — Vals AI harness (opens in a new tab)Archived copy (opens in a new tab)AccuracyReported Benchmark leaderboardVals AIIndependentRecord for Claude Sonnet 4 on MMLU-Pro Original — Vals AI harness
MMMU-Pro Vals AI harnessConfiguration: Vals AI harness72.4%vals.ai for Claude Sonnet 4 on MMMU-Pro Vals AI harness (opens in a new tab)Archived copy (opens in a new tab)MMMU-Pro accuracySetting: Standard (4 options)Reported Benchmark leaderboardVals AIIndependentRecord for Claude Sonnet 4 on MMMU-Pro Vals AI harness
MMMU-Pro Vals AI harnessReported as: thinkingConfiguration: Vals AI harness74.9%vals.ai for Claude Sonnet 4 (thinking) on MMMU-Pro Vals AI harness (opens in a new tab)Archived copy (opens in a new tab)MMMU-Pro accuracySetting: Standard (4 options)Reported Benchmark leaderboardVals AIIndependentRecord for Claude Sonnet 4 on MMMU-Pro Vals AI harness
OSWorld Verified — 100 stepsConfiguration: 100-step limit41.4%os-world.github.io for Claude Sonnet 4 on OSWorld Verified — 100 steps (opens in a new tab)Archived copy (opens in a new tab)OSWorld-Verified success rateReported Benchmark leaderboardOSWorldIndependentXLang LabRecord for Claude Sonnet 4 on OSWorld Verified — 100 steps
OSWorld Verified — 15 stepsConfiguration: 15-step limit31.2%os-world.github.io for Claude Sonnet 4 on OSWorld Verified — 15 steps (opens in a new tab)Archived copy (opens in a new tab)OSWorld-Verified success rateReported Benchmark leaderboardOSWorldIndependentXLang LabRecord for Claude Sonnet 4 on OSWorld Verified — 15 steps
OSWorld Verified — 50 stepsConfiguration: 50-step limit43.9%os-world.github.io for Claude Sonnet 4 on OSWorld Verified — 50 steps (opens in a new tab)Archived copy (opens in a new tab)OSWorld-Verified success rateReported Benchmark leaderboardOSWorldIndependentXLang LabRecord for Claude Sonnet 4 on OSWorld Verified — 50 steps
SWE-bench VerifiedEffort: NoneReported as: standard (no extended thinking)72.7%anthropic.com for Claude Sonnet 4 (standard (no extended thinking)) on SWE-bench Verified (opens in a new tab)Resolved TasksReported Launch postAnthropicSelf-reportedRecord for Claude Sonnet 4 on SWE-bench Verified
SWE-bench Verified — mini-swe-agentConfiguration: mini-swe-agent harness64.9%github.com for Claude Sonnet 4 on SWE-bench Verified — mini-swe-agent (opens in a new tab)Archived copy (opens in a new tab)Resolved TasksReported Benchmark repositorySWE-benchIndependentSWE-bench TeamRecord for Claude Sonnet 4 on SWE-bench Verified — mini-swe-agent
SWE-bench Pro Private — Scale AI SWE-AgentConfiguration: SWE-Agent harness, 250 turns, uncapped cost9.1%labs.scale.com for Claude Sonnet 4 on SWE-bench Pro Private — Scale AI SWE-Agent (opens in a new tab)Archived copy (opens in a new tab)Public tasks resolvedReported Benchmark leaderboardScale AIIndependentRecord for Claude Sonnet 4 on SWE-bench Pro Private — Scale AI SWE-Agent
SWE-bench Pro Public — Scale AI SWE-AgentConfiguration: SWE-Agent harness, 250 turns, uncapped cost42.7%labs.scale.com for Claude Sonnet 4 on SWE-bench Pro Public — Scale AI SWE-Agent (opens in a new tab)Archived copy (opens in a new tab)Public tasks resolvedReported Benchmark leaderboardScale AIIndependentRecord for Claude Sonnet 4 on SWE-bench Pro Public — Scale AI SWE-Agent
Toolathlon Original29.9%toolathlon.xyz for Claude Sonnet 4 on Toolathlon Original (opens in a new tab)Archived copy (opens in a new tab)Tasks completedReported Benchmark leaderboardToolathlonIndependentToolathlon AuthorsRecord for Claude Sonnet 4 on Toolathlon Original