Compare / Llama 3.1 8B Instruct vs GPT-4.1

Llama 3.1 8B Instruct 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.

Llama 3.1 8B Instruct
Meta · profile
GPT-4.1
OpenAI · profile
Results compared
17
Shared sources
3
BenchmarkMetricLlama 3.1 8B InstructGPT-4.1
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,2101,415
Arena (formerly LMArena)reportedCreative writingArena rating 1,1771,402
Arena (formerly LMArena)reportedExpert promptsArena rating 1,1981,405
Arena (formerly LMArena)reportedInstruction followingArena rating 1,1911,403
Arena (formerly LMArena)reportedOverallArena rating 1,2111,415
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,1821,405
Berkeley Function Calling Leaderboard (BFCL) V4reportedIrrelevance detection% correct 42.7%86.5%
Berkeley Function Calling Leaderboard (BFCL) V4reportedMemory% correct 10.8%23.9%
Berkeley Function Calling Leaderboard (BFCL) V4reportedMulti-turn tasks% correct 11.1%38.9%
Berkeley Function Calling Leaderboard (BFCL) V4reportedOverall accuracy% correct 25.8%54.0%
Berkeley Function Calling Leaderboard (BFCL) V4reportedRelevance detection% correct 93.8%100.0%
Berkeley Function Calling Leaderboard (BFCL) V4reportedSingle-turn calls (curated)% correct 84.0%88.7%
Berkeley Function Calling Leaderboard (BFCL) V4reportedSingle-turn calls (user-contributed)% correct 70.8%78.9%
Berkeley Function Calling Leaderboard (BFCL) V4reportedWeb search% correct 3.0%68.0%
UGI LeaderboardreportedRequested-length error% off the requested word count 20.0%6.0%
UGI LeaderboardreportedStyle adherencescore from 0 to 1 0.340.33
UGI LeaderboardreportedWriting scorescore out of 100 24.867.5

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.