CanMT - Japanese to English: leaderboard
Metric: Translation score (1-7) on the 116 Japanese-to-English CanMT sentences containing culture-specific items, GPT-5-nano judge with the reference translation scoring contextual accuracy, cultural adaptation, functional equivalence, fidelity and naturalness on 7-point Likert scales, averaged per sentence; default translation prompt; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4 | 5.18 |
| 2 | Grok 4.1 | 5.04 |
| 3 | GPT-4o | 5.03 |
| 4 | DeepSeek R1 | 4.97 |
| 5 | Gemini 2.5 Flash Lite | 4.96 |
| 6 | DeepSeek V3.2 | 4.88 |
| 7 | Qwen 2.5 72B Instruct | 4.67 |
| 8 | Qwen 2.5 32B Instruct | 4.6 |
| 9 | Llama 3.3 70B Instruct | 4.44 |
| 10 | Qwen 3 32B (Non-reasoning) | 4.39 |
| 11 | Qwen 2.5 14B Instruct | 4.32 |
| 12 | Qwen 3 14B (Non-reasoning) | 4.24 |
| 13 | Qwen 3 8B (Non-reasoning) | 4.02 |
| 14 | Qwen 2.5 7B Instruct | 3.82 |
| 15 | Qwen 3 4B (Non-reasoning) | 3.52 |
Interactive version: theaggregate.ai/benchmark?slug=canmt-japanese-to-english · How It Works · Data refreshed daily, snapshot 2026-10-07.