Moonshot & Alibaba release open-source models, challenging US AI lead

Chinese AI companies Moonshot AI and Alibaba have released new models to close the gap with US leaders. On Friday, Moonshot launched Kimi K3 with 2.8 trillion parameters, claiming performance second only to OpenAI's GPT-5.6 Sol and Anthropic's Claude Fable 5. Alibaba followed with Qwen3.8 (2.4T parameters), claiming it's "second only to Fable 5." Both are fully open-source, allowing free download, modification, and building, contrasting with US closed-source approaches. Moonshot will release full weights on July 27, while Alibaba says "open source coming soon." For developers, these models offer low-cost alternatives; for businesses reliant on US closed-source APIs, they provide competitive options. Their actual performance awaits independent testing, but they continue the trend of Chinese open-source models (like DeepSeek) disrupting US advantages. Despite US export controls restricting chip access, Chinese teams are overcoming hardware limits through algorithmic and engineering innovation. Regardless of final performance, these models intensify US-China tech competition and may influence policies.