Gemma 3 27B benchmark results
7 results on 7 benchmarks from 4 sources.
Also listed as: gemma-3-27b-it · Gemma-3-27b-it (Prompt) · Gemma-3-27B-it
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1365 | 198 | Overall, style control on; 46672 votes; 95% CI 1362–1369“gemma-3-27b-it” | Official leaderboard | Oct 8, 2026 |
| LMArena Vision | 1160 | 105 | Overall, style control on; 16989 votes; 95% CI 1152–1168“gemma-3-27b-it” | Official leaderboard | Oct 2, 2026 |
| GPQA Diamond | 47.7% | 214 | Epoch AI run, accuracy | Independent run | Aug 27, 2026 |
| MMLU-Pro | 67.5% | 110 | Overall accuracy; data source: Self-Reported; 27B“Gemma-3-27B-it” | Vendor-reported | Mar 11, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 22.5% | 203 | Epoch AI run, OTIS Mock AIME 2024–2025 | Independent run | Aug 27, 2026 |
| MATH Level 5 | 74.0% | 39 | Epoch AI run, accuracy | Independent run | Mar 13, 2025 |
| BFCL v4 (function calling) | 29.5% | 69 | Overall accuracy; prompt-based“Gemma-3-27b-it (Prompt)” | Official leaderboard | Oct 17, 2026 |
FAQ
Where do these Gemma 3 27B scores come from?
From Epoch AI Benchmarking Hub, Berkeley Function Calling Leaderboard (BFCL), LMArena leaderboard dataset, and MMLU-Pro leaderboard (TIGER-Lab). Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.