Compare / Gemini 2.5 Flash vs GPT-4.1

Gemini 2.5 Flash vs GPT-4.1

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

Gemini 2.5 Flash
Google · profile
GPT-4.1
OpenAI · profile
Results compared
17
Shared sources
3
BenchmarkMetricGemini 2.5 FlashGPT-4.1
Arena (formerly LMArena)reportedOverallArena rating 1,4371,408
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,3991,415
Arena (formerly LMArena)reportedCreative writingArena rating 1,3941,402
Arena (formerly LMArena)reportedExpert promptsArena rating 1,4241,405
Arena (formerly LMArena)reportedInstruction followingArena rating 1,4001,403
Arena (formerly LMArena)reportedOverallArena rating 1,4091,415
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,4021,405
Berkeley Function Calling Leaderboard (BFCL) V4reportedIrrelevance detection% correct 93.7%86.5%
Berkeley Function Calling Leaderboard (BFCL) V4reportedMemory% correct 41.3%23.9%
Berkeley Function Calling Leaderboard (BFCL) V4reportedMulti-turn tasks% correct 36.2%38.9%
Berkeley Function Calling Leaderboard (BFCL) V4reportedOverall accuracy% correct 56.2%54.0%
Berkeley Function Calling Leaderboard (BFCL) V4reportedRelevance detection% correct 75.0%100.0%
Berkeley Function Calling Leaderboard (BFCL) V4reportedSingle-turn calls (curated)% correct 88.1%88.7%
Berkeley Function Calling Leaderboard (BFCL) V4reportedSingle-turn calls (user-contributed)% correct 78.2%78.9%
Berkeley Function Calling Leaderboard (BFCL) V4reportedWeb search% correct 62.0%68.0%
Vectara Hallucination LeaderboardreportedAnswer rate% of documents 99.0%99.9%
Vectara Hallucination LeaderboardreportedHallucination rate% of summaries 7.8%5.6%

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.