Kimi K3 Is So Popular Moonshot Had to Stop Selling It
Moonshot paused new Kimi K3 subscriptions after 48 hours of demand pushed its GPU clusters to the edge.
Moonshot AI stopped taking new subscribers for Kimi K3 on July 19 — three days after launch. Requests over the first 48 hours blew past projections and pushed the company's compute clusters near capacity. Existing subscribers keep their access; new spots reopen in batches as capacity comes online.
The crush is earned. K3 is a 2.8 trillion parameter open-weight model with a 1M-token context window, and it landed at No. 1 on Arena.AI's Frontend Code Arena (1,679), ahead of Claude Fable 5 and GPT-5.6 Sol. On GDPval-AA v2 it placed third at 1,687, behind only Fable 5 Max and GPT-5.6 Sol Max. Pricing is aggressive: $3/M input, $15/M output, flat across the whole context window.
Moonshot also split memberships into two tiers — one for Kimi Web/App/Work, a separate Kimi Code plan for programming. Full weights drop July 27, which is when the compute math gets interesting for everyone else.
Why it matters: a Chinese lab hitting a demand wall it can't buy its way out of is the clearest signal yet that export controls now bite on inference, not just training.
Sources
- Kimi K3 developer suspends new subscriptions amid compute constraints scmp.com
- China's Moonshot AI releases Kimi K3, the largest open-source model ever venturebeat.com
Written by an AI pipeline from the sources above. How it works.
Feed, daily deep-dive and bytes — readable offline, with push alerts for the topics you follow.