Compare / DeepSeek V4 Flash 0423 vs Gemini 3.5 Flash

DeepSeek V4 Flash 0423 vs Gemini 3.5 Flash

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

DeepSeek V4 Flash 0423
DeepSeek · profile
Gemini 3.5 Flash
Google · profile
Results compared
19
Shared sources
3
BenchmarkMetricDeepSeek V4 Flash 0423Gemini 3.5 Flash
Arena (formerly LMArena)reportedOverallArena rating 1,4391,476
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,4391,468
Arena (formerly LMArena)reportedCreative writingArena rating 1,4081,466
Arena (formerly LMArena)reportedExpert promptsArena rating 1,4571,491
Arena (formerly LMArena)reportedInstruction followingArena rating 1,4281,465
Arena (formerly LMArena)reportedOverallArena rating 1,4361,477
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,4161,472
Artificial AnalysisreportedArtificial Analysis Coding Indexindex score 69.170.1
Artificial AnalysisreportedArtificial Analysis Intelligence Indexindex score 34.333.6
Artificial AnalysisreportedGPQA Diamond% of questions 90.8%92.2%
Artificial AnalysisreportedHumanity's Last Exam% of questions 38.6%42.7%
Artificial AnalysisreportedLong-context reasoning (AA-LCR)% of questions 79.7%74.3%
Artificial AnalysisreportedSciCode% of problems 50.3%53.9%
Artificial AnalysisreportedTerminal-Bench 2.1% of tasks 78.7%78.7%
Artificial AnalysisreportedTerminal-Bench 4.0% of tasks 12.1%6.6%
Artificial AnalysisreportedΤ-bench banking% of tasks 39.4%32.2%
UGI LeaderboardreportedRequested-length error% off the requested word count 12.0%9.0%
UGI LeaderboardreportedStyle adherencescore from 0 to 1 0.360.35
UGI LeaderboardreportedWriting scorescore out of 100 54.672.5

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.