Releases

Moonshot published the full Kimi K3 weights and implementation resources.

Kimi K3 launched with native vision, agentic capabilities and a one-million-token context window.

Moonshot introduced Kimi K3, its 2.8T-parameter native multimodal frontier model family.

Moonshot previewed an intermediate K2-generation model before Kimi K3.

Moonshot released Kimi K2.6 with improved agent capabilities and published model weights.

Moonshot published its distributed model-checkpointing engine.

Moonshot published optimized kernels for Kimi Delta Attention.

Moonshot released its expert-parallel communication library for large MoE training and inference.

Moonshot released Kimi K2.5, adding native multimodal capabilities and stronger agentic coding.

Moonshot released the reasoning-focused K2 Thinking model for long-horizon agentic tasks.

Moonshot published a 48B Kimi Linear checkpoint.

Moonshot released a K2 checkpoint with improved coding, instruction following and long-context performance.

Moonshot published the post-trained Kimi K2 Instruct checkpoint for agentic and coding tasks.

Moonshot published the base pretrained Kimi K2 checkpoint.

Moonshot introduced Kimi K2, a one-trillion-parameter mixture-of-experts model with 32B active parameters.

Moonshot released a 72B software-engineering model optimized using reinforcement learning.

Moonshot published a seven-billion-parameter formal theorem-proving model.

Moonshot published the seven-billion-parameter Kimi-Audio checkpoint.

Moonshot introduced an open audio-language model family covering audio understanding and generation.

Moonshot published the reasoning-oriented Kimi-VL A3B Thinking checkpoint.

Moonshot published the instruction-tuned Kimi-VL A3B checkpoint.

Moonshot introduced Kimi-VL, an open-weight mixture-of-experts vision-language family.

Moonshot released a preview checkpoint for formal mathematical theorem proving.

Moonshot released Moonlight, a 16B MoE model used to evaluate Muon optimization at scale.

Moonshot released Kimi k1.5, a multimodal reasoning model trained with reinforcement learning.

Moonshot introduced its open k-series reasoning work through the Kimi k1.5 release.

Moonshot exposed a 128K-context text model through its API.

Moonshot exposed a 32K-context text model through its API.

Moonshot exposed an 8K-context text model through its API.