gemma-3-4B-pt Base: benchmark results
Google's pretrained 4B Gemma 3 base checkpoint (March 2025), multimodal with a SigLIP vision encoder and 128K context. Provider: Google. Released 2025-03-12. Access: Open.
Aggregate Rating 1423 ± 14, rank #1134 of 1656 rated models, from 644 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval Slovak NLU - CSFD Sentiment SK | 65.92 | Matthews correlation coefficient (%) | 99 |
| EuroEval Czech NLU - CSFD Sentiment | 68.79 | Matthews correlation coefficient (%) | 98.2 |
| Open Arabic LLM - Aratrust Offensive | 95.65 | Accuracy (%) | 97.2 |
| LA Leaderboard - EusProficiency Basque | 33.53 | Accuracy (%) | 94.1 |
| LA Leaderboard - QNLI Basque | 57.98 | Accuracy (%) | 94.1 |
| LA Leaderboard - XCOPA Basque | 58.8 | Accuracy (%) | 94.1 |
| LA Leaderboard - xStoryCloze Basque | 59.43 | Accuracy (%) | 94.1 |
| EuroEval Finnish NLU - Scandisent FI | 92.64 | Matthews correlation coefficient (%) | 93.8 |
| EuroEval Ukrainian NLU - Cross Domain UK Reviews | 62.56 | Matthews correlation coefficient (%) | 93.8 |
| LA Leaderboard - CoQCat | 74.37 | F1 (%) | 93.8 |
| LA Leaderboard - BHTCv2 Basque | 38.24 | F1 (%) | 93.4 |
| EuroEval Swedish NLU - Swerec | 79.23 | Matthews correlation coefficient (%) | 92 |
Interactive version: theaggregate.ai/model?slug=gemma-3-4b-pt-base · How It Works · Data refreshed daily, snapshot 2026-10-20.