gemma-3n-E2B-it Instruct: benchmark results
Google's Gemma 3n E2B Instruct, a smaller instruction-tuned, image-text-to-text model derived from Gemma 3n E4B Instruct. Provider: Google. Released 2025-06-26. Access: Open.
Aggregate Rating 1374 ± 37, rank #1322 of 1656 rated models, from 59 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LLM Stats (ECLeKTic) | 2.5 | Score (%) | 41.7 |
| AA Omniscience - Software Engineering (SWE) - Dart | 12 | Accuracy (%) | 27.5 |
| AA Omniscience - Software Engineering (SWE) - Kotlin | 10 | Accuracy (%) | 22.4 |
| AA CritPt | 0 | Accuracy (%) | 22 |
| AA Humanity's Last Exam | 4.17 | Accuracy (%) | 15.1 |
| CritPt | 0 | Accuracy (imported) | 14.2 |
| AA Omniscience - Software Engineering (SWE) - Go | 6 | Accuracy (%) | 14 |
| AA AIME 2025 | 10.33 | Accuracy (%) | 13.5 |
| AA MATH-500 | 69.13 | Accuracy (%) | 13.3 |
| AA Terminal-Bench Hard | 0.76 | Accuracy (%) | 12.8 |
| LLM Stats (HiddenMath) | 27.7 | Score (%) | 12.5 |
| AA Omniscience - Software Engineering (SWE) - Python | 7.5 | Accuracy (%) | 10.9 |
Interactive version: theaggregate.ai/model?slug=gemma-3n-e2b-it-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.