Kimi K2 Thinking outperforms US rivals

Moonshot AI has launched Kimi K2 Thinking, a one-trillion-parameter reasoning model that reportedly outperforms OpenAI’s GPT-5 and Anthropic’s Claude Sonnet 4.5 on key benchmarks. The Beijing-based firm said the open-source model’s MoE design enhances reasoning and agentic capabilities while providing API access up to ten times cheaper than US counterparts.