PaliBench - chrF++: leaderboard
Metric: chrF++ (0-100), character n-gram F-score with word bigrams (sacrebleu), 1,700 Pali canon passages translated into English, each scored against three human reference translations (Bodhi, Sujato, Thanissaro), corpus level, models through OpenRouter; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 68.5 |
| 2 | Gemini 3 Flash (Preview) | 65.9 |
| 3 | Claude Opus 4.5 | 65.6 |
| 4 | DeepSeek V3.2 | 64.1 |
| 5 | Kimi K2.5 | 61.2 |
| 6 | GPT-5.2 | 59.1 |
| 7 | GLM-4.7 | 57.1 |
| 8 | Qwen 3 235B A22B 2507 Instruct | 55.7 |
| 9 | Grok 4.1 Fast | 54.8 |
| 10 | Llama 3.3 70B Instruct | 48.1 |
Interactive version: theaggregate.ai/benchmark?slug=palibench-chrf-plus-plus · How It Works · Data refreshed daily, snapshot 2026-10-07.