Compare / GPT-4.1 vs GPT-5 Nano
GPT-4.1 vs GPT-5 Nano
16 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | GPT-4.1 | GPT-5 Nano |
|---|---|---|---|
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,415 | 1,339 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,402 | 1,247 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,405 | 1,354 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,403 | 1,327 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,415 | 1,338 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,405 | 1,289 |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Irrelevance detection% correct | 86.5% | 89.1% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Memory% correct | 23.9% | 24.7% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Multi-turn tasks% correct | 38.9% | 34.5% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Overall accuracy% correct | 54.0% | 51.5% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Relevance detection% correct | 100.0% | 93.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (curated)% correct | 88.7% | 80.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (user-contributed)% correct | 78.9% | 70.7% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Web search% correct | 68.0% | 72.5% |
| Vectara Hallucination Leaderboardreported | Answer rate% of documents | 99.9% | 100.0% |
| Vectara Hallucination Leaderboardreported | Hallucination rate% of summaries | 5.6% | 10.5% |
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.