Gemma 3 12B — benchmark results
Google Gemma 3 12B model row. Provider: Google. Released 2025-03-12. Access: Open.
Unified ELO 1464 ± 5, rank #934 of 1776 rated models, from 574 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM MMT-Bench - Traffic Anomaly Detection | 75 | Score (%) | 99.5 |
| OpenVLM MMT-Bench - Scene Recognition | 85 | Score (%) | 97.3 |
| OpenVLM LLaVA-Bench - Complex | 108.1 | Score | 97 |
| OpenVLM LLaVA-Bench - Detail | 93.5 | Score | 96.1 |
| OpenVLM MMT-Bench - MeViS | 50 | Score (%) | 96.1 |
| OpenVLM MMT-Bench - Logo and Brand Recognition | 100 | Score (%) | 94.4 |
| OpenVLM MMT-Bench - Shape Recognition | 95 | Score (%) | 93.7 |
| OpenVLM MMT-Bench - OCR | 77.5 | Score (%) | 93.4 |
| OpenVLM MMT-Bench - Color Assimilation | 75 | Score (%) | 93 |
| OpenVLM MMT-Bench - Facial Expression Change Recognition | 95 | Score (%) | 93 |
| OpenVLM MMT-Bench - Jigsaw Puzzle Solving | 40 | Score (%) | 92.7 |
| OpenVLM MMT-Bench - Food Recognition | 100 | Score (%) | 92.5 |
Interactive version: theaggregate.ai/model?slug=gemma-3-12b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.