Compare / Phi 4 vs Mistral Medium 3

Phi 4 vs Mistral Medium 3

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

Phi 4
Microsoft · profile
Mistral Medium 3
Mistral · profile
Results compared
17
Shared sources
3
BenchmarkMetricPhi 4Mistral Medium 3
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,2561,387
Arena (formerly LMArena)reportedCreative writingArena rating 1,2101,364
Arena (formerly LMArena)reportedExpert promptsArena rating 1,2681,379
Arena (formerly LMArena)reportedInstruction followingArena rating 1,2451,366
Arena (formerly LMArena)reportedOverallArena rating 1,2561,387
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,2291,370
Berkeley Function Calling Leaderboard (BFCL) V4reportedIrrelevance detection% correct 87.5%92.0%
Berkeley Function Calling Leaderboard (BFCL) V4reportedMemory% correct 24.7%23.0%
Berkeley Function Calling Leaderboard (BFCL) V4reportedMulti-turn tasks% correct 3.9%10.8%
Berkeley Function Calling Leaderboard (BFCL) V4reportedOverall accuracy% correct 28.8%37.7%
Berkeley Function Calling Leaderboard (BFCL) V4reportedRelevance detection% correct 50.0%75.0%
Berkeley Function Calling Leaderboard (BFCL) V4reportedSingle-turn calls (curated)% correct 69.6%85.3%
Berkeley Function Calling Leaderboard (BFCL) V4reportedSingle-turn calls (user-contributed)% correct 60.7%68.0%
Berkeley Function Calling Leaderboard (BFCL) V4reportedWeb search% correct 4.5%39.0%
UGI LeaderboardreportedRequested-length error% off the requested word count 19.0%20.0%
UGI LeaderboardreportedStyle adherencescore from 0 to 1 0.340.35
UGI LeaderboardreportedWriting scorescore out of 100 25.732.6

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.