CanMT - Chinese to Russian: leaderboard

Metric: Translation score (1-7) on the 95 Chinese-to-Russian 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

#ModelScore
1Grok 4.15.3
2DeepSeek R15.22
3DeepSeek V3.25
4Gemini 2.5 Flash Lite4.97
5GPT-44.88
6Llama 3.3 70B Instruct4.8
7GPT-4o4.75
8Qwen 2.5 72B Instruct4.7
9Qwen 3 32B (Non-reasoning)4.34
10Qwen 3 14B (Non-reasoning)4.27
11Qwen 2.5 32B Instruct4.05
12Qwen 3 8B (Non-reasoning)4.05
13Mixtral 8x7B Instruct (v0.1)3.73
14Qwen 2.5 14B Instruct3.66
15Qwen 3 4B (Non-reasoning)3.46

Interactive version: theaggregate.ai/benchmark?slug=canmt-chinese-to-russian · How It Works · Data refreshed daily, snapshot 2026-10-07.