Moonshot AI launches 2.8T Kimi K3
Moonshot AI has launched Kimi K3, a flagship 2.8 trillion-parameter open-weight multimodal reasoning model featuring a 1-million-token context window. The model uses Kimi Delta Attention and Attention Residuals to optimize performance for long-horizon coding, complex knowledge work, and deep reasoning.
Moonshot AI's Kimi K3 release demonstrates that Chinese AI labs are not just keeping pace with Western counterparts but are leading the deployment of trillion-parameter scale open-weight architectures.
* At 2.8 trillion parameters, Kimi K3 is currently the largest open-weight model, significantly closing the gap between proprietary and open-source systems.
* The incorporation of hybrid linear attention (Kimi Delta Attention) enables faster decoding and better resource efficiency for ultra-long context tasks.
* Strong capabilities in frontend coding and multi-step reasoning position the model as a highly competitive option for developer and agentic workflows.
* While it may still trail top-tier closed-source systems like GPT-5 class models in overall capability, its open accessibility represents a major milestone for open-source AI.
DISCOVERED
2h ago
2026-07-17
PUBLISHED
2h ago
2026-07-17
RELEVANCE
AUTHOR
LinusEkenstam
