Chinese AI Lab Releases Free Model That Claims to Outperform GPT-5 for Just $4.6 Million

Nov 11, 2025
ZDNET
Article image for Chinese AI Lab Releases Free Model That Claims to Outperform GPT-5 for Just $4.6 Million

Summary

Chinese AI lab Moonshot releases Kimi K2 Thinking, a free open-source reasoning model that reportedly outperforms GPT-5 and Claude Sonnet 4.5 on major benchmarks while costing just $4.6 million to train, potentially disrupting the billion-dollar AI market and intensifying US-China competition.

Key Points

  • Chinese AI lab Moonshot releases Kimi K2 Thinking, a free open-source reasoning model that claims to outperform GPT-5 and Claude Sonnet 4.5 on key benchmarks including Humanity's Last Exam and BrowseComp
  • The model costs only $4.6 million to train compared to billions spent by US AI labs, raising questions about the commercial viability of expensive proprietary AI subscriptions
  • The release intensifies the US-China AI competition as businesses gain access to advanced agentic capabilities at no cost, potentially disrupting the current AI market structure

Tags

Read Original Article