Compare / DeepSeek V3 vs Gemma 3 4B
DeepSeek V3 vs Gemma 3 4B
11 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | DeepSeek V3 | Gemma 3 4B |
|---|---|---|---|
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,350 | 1,330 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,349 | 1,274 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,348 | 1,260 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,343 | 1,267 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,358 | 1,303 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,356 | 1,298 |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 17.0% | 9.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.37 | 0.29 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 43.2 | 24.0 |
| Vectara Hallucination Leaderboardreported | Answer rate% of documents | 97.5% | 67.3% |
| Vectara Hallucination Leaderboardreported | Hallucination rate% of summaries | 6.1% | 6.4% |
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.