Moonshot AI introduces Kimi K3, an open-source Mixture-of-Experts model featuring 2.8 trillion total parameters and 104 billion activated parameters per token. The architecture utilizes native vision capabilities and supports a 1-million-token context window, enabled by Kimi Delta Attention and Stable LatentMoE.

  • Kimi K3 achieves approximately 2.5x improvement in overall scaling efficiency over its predecessor, Kimi K2.
  • The model employs Stable LatentMoE to effectively activate only 16 of 896 routed experts per token.
  • Post-training includes reinforcement learning across general, agentic, and coding domains with multiple reasoning-effort levels.
  • Extensive evaluations show frontier-level performance in long-horizon coding, agentic, knowledge, reasoning, and vision tasks.
  • While trailing Claude Fable 5 and GPT-5.6 Sol, Kimi K3 outperforms other open and proprietary models in the evaluation suite.

Moonshot AI releases the full model weights to facilitate future research and accelerate the broader deployment of frontier intelligence.