Monthly Archive October 18, 2026

Important Notice for Overseas Telegraphic Transfer (TT):

1. Bank Address: Queen’s Road Central, Hong Kong
2. Account Currency: USD / HKD / EUR
3. Payment Reference: Please write Order #{order_number} in the bank transfer note/memo so we can process your order promptly.
4. Bank Charges: Please select OUR (sender pays all transaction fees) to ensure full amount is received.

In the following examples, both source_lang and target_lang should use the full language names. Chinese names should be used in Chinese prompts, and English names should be used in English prompts.

• 2026.5.21 We open-sourced Hy-MT2-1.8B/Hy-MT2-7B/Hy-MT2-30B-A3B/IFMTBench on HuggingFace and ModelScope.
• 2025.12.30 We open-sourced HY-MT1.5-1.8B and HY-MT1.5-7B on HuggingFace and ModelScope.
• 2025.9.1 We open-sourced Hunyuan-MT-7B and Hunyuan-MT-Chimera-7B on HuggingFace and ModelScope.

Now, Tencent Hy is officially partnering with WMT26 for the “Video Subtitle Translation Task” (https://www2.statmt.org/wmt26/video-subtitle-translation.html). Participants who use the Hy-MT model series to compete in the “General Machine Translation Task” (https://www2.statmt.org/wmt26/translation-task.html) and the “Video Subtitle Translation Task” will have the chance to win special awards sponsored by Hunyuan. We sincerely invite everyone to participate and jointly push the boundaries of machine translation technology!

In this release, we also open-source IFMTBench, a benchmark for evaluating translation instruction-following capabilities.

1. Structure Preservation: You MUST preserve the original {format_type} data structure, nesting, hierarchy, and indentation exactly as they are.
2. Selective Translation: Translate ONLY the visible, user-facing text content/values.
3. Strict Non-Translation: NEVER translate or alter code tags, keys, properties, object names, or variable placeholders. Leave them exactly in their original English/code form.

Hy-MT2 is a family of “fast-thinking” multilingual translation models designed for complex real-world scenarios. It includes three model sizes: 1.8B, 7B, and 30B-A3B (MoE), all of which support translation among 33 languages and effectively follow translation instructions in multiple languages. For on-device deployment, AngelSlim 1.25-bit extreme quantization reduces the storage requirement of the 1.8B model to only 440 MB and improves inference speed by 1.5x. Multi-dimensional evaluations show that Hy-MT2 delivers outstanding performance across general, real-world business, domain-specific, and instruction-following translation tasks. The 7B and 30B-A3B models outperform open-source models such as DeepSeek-V4-Pro and Kimi K2.6 in fast-thinking mode, while the lightweight 1.8B model also surpasses mainstream commercial APIs from providers such as Microsoft and Doubao overall.