Gemma 3 27B: benchmark results
Google's open-weight Gemma 3 27B, the flagship multimodal model of the Gemma 3 family (March 2025). Provider: Google. Released 2025-03-12. Access: Open.
Aggregate Rating 1553 ± 4, rank #607 of 1656 rated models, from 1213 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AICD Bench - Robust Binary Classification (Zero-Shot) | 58.45 | Macro-F1 (%; human-written vs machine-generated code under u | 100 |
| BloomBench - Arabic - Apply | 84 | Accuracy (%) on the 499 Arabic BloomBench items at the Bloom | 100 |
| BloomBench - Arabic - Understand | 93 | Accuracy (%) on the 1592 Arabic BloomBench items at the Bloo | 100 |
| DNBench - BIRD - Complex - Few-shot | 34.1 | DNB-Score (0-100) | 100 |
| DNBench - BIRD - Complex - Zero-shot | 35.6 | DNB-Score (0-100) | 100 |
| DNBench - BIRD - Real World - Few-shot | 22.2 | DNB-Score (0-100) | 100 |
| DNBench - Complex - Few-shot | 36.6 | DNB-Score (0-100) | 100 |
| DNBench - Component - Column F1 | 92.1 | Component Score (0-100) | 100 |
| DNBench - Component - Lossless Join | 63.4 | Component Score (0-100) | 100 |
| DNBench - Real World - Few-shot | 22 | DNB-Score (0-100) | 100 |
| DNBench - Spider - Complex - Few-shot | 37.6 | DNB-Score (0-100) | 100 |
| DNBench - Spider - Real World - Few-shot | 21.8 | DNB-Score (0-100) | 100 |
Interactive version: theaggregate.ai/model?slug=gemma-3-27b · How It Works · Data refreshed daily, snapshot 2026-10-20.