Compare / Claude Sonnet 4.5 vs Mistral Small 3.2 24B
Claude Sonnet 4.5 vs Mistral Small 3.2 24B
17 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Claude Sonnet 4.5 | Mistral Small 3.2 24B |
|---|---|---|---|
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,461 | 1,363 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,454 | 1,323 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,498 | 1,335 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,464 | 1,338 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,457 | 1,357 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,454 | 1,329 |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Irrelevance detection% correct | 95.0% | 87.9% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Memory% correct | 65.0% | 18.1% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Multi-turn tasks% correct | 61.4% | 14.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Overall accuracy% correct | 73.2% | 37.1% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Relevance detection% correct | 68.8% | 93.8% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (curated)% correct | 88.7% | 89.7% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (user-contributed)% correct | 81.1% | 79.0% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Web search% correct | 81.0% | 31.0% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 5.0% | 17.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.41 | 0.35 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 66.9 | 36.3 |
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.