China's Open-Source Models Top Global Rankings: Kimi, DeepSeek, and Qwen Lead the Way

中国开源模型领跑全球：Kimi、DeepSeek、Qwen登顶榜单

China's self-developed open-source AI models are shining on the international stage.

中国自研的开源大模型在国际舞台上大放异彩。

Kimi K2, DeepSeek R1, and Qwen3 have taken the top three spots on the global leaderboard, surpassing Google's Gemma3 and Meta's Llama4.

Kimi K2、DeepSeek R1 和 Qwen3 分别荣登全球排行榜前三名，超越谷歌的Gemma3和Meta的Llama4。

The ranking is based on dynamic blind testing votes from developers worldwide, lending it strong credibility.

该榜单由全球开发者采用动态盲测投票方式评选，具有高度权威性。

NVIDIA CEO Jensen Huang praised these three models as world-leading representatives of open-source AI.

英伟达CEO黄仁勋也公开肯定这三大模型，称其为全球领先的开源代表。

This breakthrough underscores China's rising strength in core AI technologies.

这一突破彰显中国在人工智能核心技术领域的强劲竞争力。