Compare / Llama 3.1 8B Instruct vs Llama 3.3 70B Instruct
Llama 3.1 8B Instruct vs Llama 3.3 70B 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 | Llama 3.1 8B Instruct | Llama 3.3 70B Instruct |
|---|---|---|---|
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,210 | 1,304 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,177 | 1,285 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,198 | 1,298 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,191 | 1,293 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,211 | 1,318 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,182 | 1,292 |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Irrelevance detection% correct | 42.7% | 53.5% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Memory% correct | 10.8% | 8.2% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Multi-turn tasks% correct | 11.1% | 21.5% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Overall accuracy% correct | 25.8% | 31.9% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Relevance detection% correct | 93.8% | 100.0% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (curated)% correct | 84.0% | 88.0% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Single-turn calls (user-contributed)% correct | 70.8% | 76.6% |
| Berkeley Function Calling Leaderboard (BFCL) V4reported | Web search% correct | 3.0% | 10.0% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 20.0% | 10.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.34 | 0.37 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 24.8 | 26.2 |
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.