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.
Unified ELO 1513 ± 1, rank #631 of 1392 rated models, from 616 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM LLaVA-Bench - Complex | 112.4 | Score | 100 |
| OpenVLM LLaVA-Bench - Detail | 102.9 | Score | 99.6 |
| OpenVLM MMT-Bench - Chemical Apparatus Recognition | 80 | Score (%) | 99 |
| OpenVLM MMMU - Energy and Power | 73.3 | Accuracy (%) | 98.8 |
| LLaVA-Bench | 103.1 | Score | 98.7 |
| OpenVLM MMT-Bench - Rock Recognition | 80 | Score (%) | 98.3 |
| OpenVLM MMMU - Electronics | 86.7 | Accuracy (%) | 98.2 |
| OpenVLM MMT-Bench - Art Design | 90.9 | Score (%) | 98.1 |
| OpenVLM MMBench V1.1 CN - Identity Reasoning | 100 | Accuracy (%) | 97.9 |
| OpenVLM MMMU - History | 83.3 | Accuracy (%) | 96.5 |
| OpenVLM MMT-Bench - DU | 86.4 | Score (%) | 96.4 |
| OpenVLM MMMU - Art Theory | 93.3 | Accuracy (%) | 96.1 |
Interactive version: theaggregate.ai/model?slug=gemma-3-27b · How It Works · Data refreshed daily, snapshot 2026-09-05.