Compare / Claude Opus 4.5 vs Llama 3.2 3B Instruct
Claude Opus 4.5 vs Llama 3.2 3B Instruct
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 Opus 4.5 | Llama 3.2 3B Instruct |
|---|---|---|---|
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,472 | 1,146 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,469 | 1,145 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,503 | 1,179 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,483 | 1,146 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,474 | 1,167 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,467 | 1,143 |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Irrelevance detection% correct | 90.8% | 52.1% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Memory% correct | 73.8% | 6.2% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Multi-turn tasks% correct | 68.4% | 4.0% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Overall accuracy% correct | 77.5% | 21.9% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Relevance detection% correct | 68.8% | 87.5% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (curated)% correct | 89.7% | 82.7% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (user-contributed)% correct | 79.8% | 58.3% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Web search% correct | 84.5% | 1.0% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 3.0% | 31.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.39 | 0.36 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 70.4 | 10.4 |
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.