Today, Moonshot AI releases the full weights for Kimi K3, its 2.8-trillion-parameter flagship model, on Hugging Face. The Beijing-based lab has been running a countdown timer on the repository page for days. Kimi K3 uses a Mixture-of-Experts architecture with 896 experts, activating just 16 per token, meaning roughly 50 billion parameters are doing the actual work during any given inference pass. It ships with a 1-million-token context window, native multimodal capabilities, and two architectural innovations called Kimi Delta Attention and Attention Residuals, which deliver up to 6.3x faster decoding in million-token contexts. The model launched via API on July 16. On the Artificial Analysis Intelligence Index, K3 scores 57, placing it fourth overall behind Claude Fable 5, GPT-5.6 Sol, and Claude Opus 4.8. K3 leads all tested models on Program Bench with a score of 77.8, wins Terminal Bench 2.1 against every model except GPT-5.6 Sol, and finished first outright on BrowseComp. Moonshot acknowledges that K3 trails Fable 5 and GPT-5.6 Sol on overall intelligence metrics. Moonshot is releasing the trained model parameters, not the training code, data, or full reproduction pipeline, under what reporting suggests will be a Modified MIT license.