Benchmarks

31 results across 12 benchmarks

Benchmark results for DeepSeek-V4-Flash
BenchmarkScoreSource
AutomationBench 1.0.6Effort: Max18.1%zapier.com for DeepSeek-V4-Flash (max) on AutomationBench 1.0.6 (opens in a new tab)Archived copy (opens in a new tab)Task pass rateReported Benchmark leaderboardZapierIndependentZapier AutomationBench TeamRecord for DeepSeek-V4-Flash on AutomationBench 1.0.6
BrowseCompEffort: High53.5%huggingface.co for DeepSeek-V4-Flash (high) on BrowseComp BrowseComp (opens in a new tab)BrowseComp AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on BrowseComp BrowseComp
BrowseCompEffort: Max73.2%huggingface.co for DeepSeek-V4-Flash (max) on BrowseComp BrowseComp (opens in a new tab)BrowseComp AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on BrowseComp BrowseComp
DeepSWE 1.1Effort: Max53.0%deepswe.datacurve.ai for DeepSeek-V4-Flash (max) on DeepSWE 1.1 (opens in a new tab)DeepSWE Pass@1Reported Benchmark leaderboardDatacurveIndependentRecord for DeepSeek-V4-Flash on DeepSWE 1.1
GPQA DiamondEffort: NoneReported as: non-thinking71.2%huggingface.co for DeepSeek-V4-Flash (non-thinking) on Graduate-Level Google-Proof Q&A Diamond (opens in a new tab)GPQA Diamond AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Graduate-Level Google-Proof Q&A Diamond
GPQA DiamondEffort: High87.4%huggingface.co for DeepSeek-V4-Flash (high) on Graduate-Level Google-Proof Q&A Diamond (opens in a new tab)GPQA Diamond AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Graduate-Level Google-Proof Q&A Diamond
GPQA DiamondEffort: Max88.1%huggingface.co for DeepSeek-V4-Flash (max) on Graduate-Level Google-Proof Q&A Diamond (opens in a new tab)GPQA Diamond AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Graduate-Level Google-Proof Q&A Diamond
Humanity's Last Exam Full set — No toolsEffort: MaxConfiguration: No tools34.8%huggingface.co for DeepSeek-V4-Flash (max) on Humanity's Last Exam Full set — No tools (opens in a new tab)Humanity's Last Exam AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Humanity's Last Exam Full set — No tools
Humanity's Last Exam Full set — With toolsEffort: MaxConfiguration: With tools45.1%huggingface.co for DeepSeek-V4-Flash (max) on Humanity's Last Exam Full set — With tools (opens in a new tab)Humanity's Last Exam AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Humanity's Last Exam Full set — With tools
MMLU-Pro OriginalEffort: NoneReported as: non-thinking83.0%huggingface.co for DeepSeek-V4-Flash (non-thinking) on MMLU-Pro Original (opens in a new tab)AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on MMLU-Pro Original
MMLU-Pro OriginalEffort: High86.4%huggingface.co for DeepSeek-V4-Flash (high) on MMLU-Pro Original (opens in a new tab)AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on MMLU-Pro Original
MMLU-Pro OriginalEffort: Max86.2%huggingface.co for DeepSeek-V4-Flash (max) on MMLU-Pro Original (opens in a new tab)AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on MMLU-Pro Original
Riemann-bench October 2026 leaderboardEffort: Max10.4%surgehq.ai for DeepSeek-V4-Flash (Max) on Riemann-bench October 2026 leaderboard (opens in a new tab)Archived copy (opens in a new tab)Riemann-bench Pass@1Reported Benchmark leaderboardSurge AIIndependentRecord for DeepSeek-V4-Flash on Riemann-bench October 2026 leaderboard
SimpleQA Verified OriginalEffort: NoneReported as: non-thinking23.1%huggingface.co for DeepSeek-V4-Flash (non-thinking) on SimpleQA Verified Original (opens in a new tab)AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SimpleQA Verified Original
SimpleQA Verified OriginalEffort: High28.9%huggingface.co for DeepSeek-V4-Flash (high) on SimpleQA Verified Original (opens in a new tab)AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SimpleQA Verified Original
SimpleQA Verified OriginalEffort: Max34.1%huggingface.co for DeepSeek-V4-Flash (max) on SimpleQA Verified Original (opens in a new tab)AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SimpleQA Verified Original
SWE-bench MultilingualEffort: NoneReported as: non-thinking69.7%huggingface.co for DeepSeek-V4-Flash (non-thinking) on SWE-bench Multilingual (opens in a new tab)Resolved TasksReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SWE-bench Multilingual
SWE-bench MultilingualEffort: High70.2%huggingface.co for DeepSeek-V4-Flash (high) on SWE-bench Multilingual (opens in a new tab)Resolved TasksReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SWE-bench Multilingual
SWE-bench MultilingualEffort: Max73.3%huggingface.co for DeepSeek-V4-Flash (max) on SWE-bench Multilingual (opens in a new tab)Resolved TasksReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SWE-bench Multilingual
SWE-bench VerifiedEffort: NoneReported as: non-thinking73.7%huggingface.co for DeepSeek-V4-Flash (non-thinking) on SWE-bench Verified (opens in a new tab)Resolved TasksReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SWE-bench Verified
SWE-bench VerifiedEffort: High78.6%huggingface.co for DeepSeek-V4-Flash (high) on SWE-bench Verified (opens in a new tab)Resolved TasksReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SWE-bench Verified
SWE-bench VerifiedEffort: Max79.0%huggingface.co for DeepSeek-V4-Flash (max) on SWE-bench Verified (opens in a new tab)Resolved TasksReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SWE-bench Verified
SWE-bench Pro PublicEffort: Max52.6%huggingface.co for DeepSeek-V4-Flash (max) on SWE-bench Pro Public (opens in a new tab)Public tasks resolvedReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on SWE-bench Pro Public
Terminal-Bench 2.0Effort: NoneReported as: non-thinking49.1%huggingface.co for DeepSeek-V4-Flash (non-thinking) on Terminal-Bench 2.0 (opens in a new tab)Terminal-Bench AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Terminal-Bench 2.0
Terminal-Bench 2.0Effort: High56.6%huggingface.co for DeepSeek-V4-Flash (high) on Terminal-Bench 2.0 (opens in a new tab)Terminal-Bench AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Terminal-Bench 2.0
Terminal-Bench 2.0Effort: Max56.9%huggingface.co for DeepSeek-V4-Flash (max) on Terminal-Bench 2.0 (opens in a new tab)Terminal-Bench AccuracyReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Terminal-Bench 2.0
Toolathlon OriginalEffort: NoneReported as: non-thinking40.7%huggingface.co for DeepSeek-V4-Flash (non-thinking) on Toolathlon Original (opens in a new tab)Tasks completedReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Toolathlon Original
Toolathlon OriginalEffort: High43.5%huggingface.co for DeepSeek-V4-Flash (high) on Toolathlon Original (opens in a new tab)Tasks completedReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Toolathlon Original
Toolathlon OriginalEffort: Max47.8%huggingface.co for DeepSeek-V4-Flash (max) on Toolathlon Original (opens in a new tab)Tasks completedReported Model cardDeepSeekSelf-reportedDeepSeek AIRecord for DeepSeek-V4-Flash on Toolathlon Original
Toolathlon OriginalEffort: Max48.2%toolathlon.xyz for DeepSeek-V4-Flash (Max) on Toolathlon Original (opens in a new tab)Archived copy (opens in a new tab)Tasks completedReported Benchmark leaderboardToolathlonIndependentToolathlon AuthorsRecord for DeepSeek-V4-Flash on Toolathlon Original
Toolathlon VerifiedEffort: Max50.9%toolathlon.xyz for DeepSeek-V4-Flash (max) on Toolathlon Verified (opens in a new tab)Archived copy (opens in a new tab)Tasks completedReported Benchmark leaderboardToolathlonIndependentToolathlon AuthorsRecord for DeepSeek-V4-Flash on Toolathlon Verified