[REVISION HISTORY]

Development and application of Kimi K3 AI model

Updated 1 time since CLSTR started tracking revisions of this situation.

What changed

2026-08-25 17:07 UTC → 2026-08-26 23:09 UTC · added ~~removed~~

Moonshot AI developed Kimi K3, a multimodal artificial intelligence model from China featuring a 2.8 trillion parameter architecture and a context window of up to one million tokens. The model was designed for research, reasoning, and information processing. Following its release, the legal AI platform Harvey launched Tenet, a proprietary language model specifically for legal tasks such as contract drafting and document review. Tenet was built by post-training on the weights of the open-source Kimi K3. While the model allows organizations to deploy technology without sharing data with the original developer, the use of a Chinese open-source base for legal professionals has raised concerns regarding data privacy and confidentiality. Moonshot AI has since entered early-stage negotiations with U.S. cloud providers Microsoft, Amazon, and Google regarding revenue-sharing agreements. These proposed deals would allow the Kimi K3 model to be hosted on platforms such as Azure, AWS, and Google Cloud. Moonshot is reportedly seeking up to 30% of the revenue generated from services related to the model, which has shown performance comparable to top-tier models like OpenAI’s GPT-5.5 and Anthropic’s Claude Opus 4.8 in independent evaluations. Significant hurdles remain in these negotiations, including unresolved issues regarding data access, revenue split mechanisms, and the auditing of token usage. These discussions take place amid heightened geopolitical tensions and scrutiny from U.S. officials concerning national security and AI hardware acquisition. Moonshot has previously denied allegations of illegal chip acquisition or model distillation.

Versions

Only revisions since CLSTR began indexing content versions appear here. Select a version to see what changed compared to the one before it.