Persian Fluency Bench - Naturalness: leaderboard
Metric: Judge Score (%, judge google/gemini-2.5-flash). Source: github.com. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Flash | 52.8 |
| 2 | Gemma 3 12B (IT) | 50.1 |
| 3 | Qwen 3 Next 80B A3B Instruct | 46.8 |
| 4 | Gemma 3 27B (IT) | 44.8 |
| 5 | Qwen 3 30B A3B 2507 Instruct | 42.9 |
| 6 | Gemma 3 4B (IT) | 42 |
| 7 | Gemma 3n E4B (IT) | 39.1 |
| 8 | ERNIE 4.5 21B A3B | 25.9 |
| 9 | granite-4.0-h-micro | 12.8 |
| 10 | Qwen 3 4B | 9.2 |
| 11 | Llama 3.2 3B Instruct | 8.7 |
| 12 | Ministral 3B | 7.4 |
| 13 | Mixtral 8x7B Instruct | 5.5 |
| 14 | LFM2 8B A1B | 3.7 |
Interactive version: theaggregate.ai/benchmark?slug=persian-fluency-bench-naturalness · How It Works · Data refreshed daily, snapshot 2026-09-20.