Gemma 4 E2B (IT) (Non-reasoning): benchmark results
Provider: Google. Released 2026-04-02. Access: Open.
Unified ELO 1416 ± 1, rank #2328 of 3078 rated models, from 28 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| FrameBench - Frame Identification - English | 74 | Accuracy (%; FrameNet candidate frames) | 36.8 |
| FrameBench - Frame Identification - Japanese | 67.8 | Accuracy (%; Japanese FrameNet candidate frames) | 32 |
| MERA v2 - NewReasoning | 24.3 | Score (%) | 26.3 |
| MERA v2 - IFHardBench | 5 | Score (%) | 25 |
| MERA v2 - SOB-Hard | 1.4 | Score (%) | 25 |
| MERA v2 - Reasoning | 22.7 | Score (%) | 21.1 |
| MERA v2 - Enantiosemy | 14.6 | Score (%) | 18.4 |
| FrameBench - English | 68.6 | Accuracy (%; mean of five prompt templates) | 15.8 |
| MERA v2 | 0.1 | Total score | 13.2 |
| MERA v2 - Agentic | 3.3 | Score (%) | 13.2 |
| MERA v2 - Characters | 29 | Score (%) | 13.2 |
| MERA v2 - Humor | 5.2 | Score (%) | 13.2 |
Interactive version: theaggregate.ai/model?slug=gemma-4-e2b-it-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-19.