Moonshot AI’s new Kimi K2 Thinking and Kimi K2 Thinking Turbo models alternate between cycles of reasoning and tool use, often making hundreds of calls, to outperform other open-weights LLMs on complex, multi-step tasks. Built as trillion-parameter mixture-of-experts models and fine-tuned at INT4 precision, they deliver strong agentic performance while running on lower-cost hardware. Learn more in The Batch: https://hubs.la/Q03VvFRw0
Document processing | Data & AI
1hKimi K2 is impressive. A lot of business use cases feels the same as GPT5 and Gemini 3. Kimi.com Chatbox is one of the few that can get answers to decent presentations. On the technical side, interleaved thinking is enlightening.