腾讯混元大模型（Hunyuan）是腾讯自研的全模态大模型家族，Hunyuan-Large 是业界首个开源万亿参数 MoE 大模型，Hunyuan-A13B 采用非 Attention 架构，2025 年开源的多模态与文生视频模型广受关注。提供 API 与开源权重，支持中文场景的对话、图像、视频生成，是中文开源社区的重要力量。 🔗 文章: /post/best-ai-apis-for-hong-kong-developers-in-2026-10-providers-compared-20260723
Tencent Hunyuan is Tencents self-developed full-modal LLM family. Hunyuan-Large was the industrys first open-source trillion-parameter MoE model, and Hunyuan-A13B uses a non-Attention architecture. Its open multimodal and text-to-video models drew wide attention in 2025, with both API and open weights for Chinese-language conversation, image and video generation. 🔗 Article: /post/best-ai-apis-for-hong-kong-developers-in-2026-10-providers-compared-20260723