Chinese Moonshot AI Released Kimi K3, an Open Model with 2.8 Trillion Parameters
July 20, 2026

IBL News | New York
Beijing-based Moonshot AI last week introduced Kimi K3, a 2.8 trillion-parameter open model that, according to the company, performs similarly to Claude Fable 5 and GPT 5.6 Sol.
The response, beyond expectations, was that the company communicated yesterday that it was pausing new subscriptions, saying that over the past two days, Kimi K3 demand had nearly reached the limit of its capacity.
To keep the experience stable, Moonshot AI said that it was splitting into two more focused plans: Kimi Membership for Kimi Web, App, and Work; and Kimi Code Membership for coding workflows.
At launch, Kimi K3 was available on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. The full model weights will be released by July 27, 2026.
Kimi K3 is the first open model to reach 2.8 trillion parameters and a 1-million-token context window. That context window means the user can feed it entire codebases, massive documents, or hours of conversation without it forgetting anything.

Kimi K3 is built on Kimi Delta Attention (KDA) and Attention Residuals (AttnRes), two architectural updates designed to improve the flow of information across sequence length and model depth.
The Kimi Delta Attention architecture delivers 6.3x faster decoding at million-token lengths. Instead of re-reading everything each time, it only processes what changed. The model also activates only 16 of 896 internal “expert” sub-networks per request, keeping compute usage lean.
“Kimi K3 has strong long-horizon coding performance. Operating with minimal human oversight, it can sustain long engineering sessions, navigate massive repositories, and orchestrate terminal tools,” said the company. “Kimi K3 excels at motion design, animation, and video editing because its native multimodal architecture understands text, images, and video within the same model.”
• Kimi K3 Agents: The latest Kimi app from the mobile app store, available on iOS, Android, and HarmonyOS, or at kimi.com
• Work with Kimi K3: The latest Kimi Work desktop app, version 3.1.0 or later, is available to download for Windows and Apple silicon Macs.
• Code with Kimi K3: Run Kimi Code in the user terminal, selecting Kimi K3 using the /model command.
• Build with the Kimi API: Visit the Kimi API Platform and select kimi-k3. Pricing is $0.30/MTok for cache-hit input, $3.00/MTok for cache-miss input, and $15.00/MTok for output.
Kimi K3 has received far more love than we expected, and our GPUs are feeling it.
Over the past 48 hours, demand has pushed close to the limits of our current capacity. To protect the experience of existing subscribers, we’re temporarily pausing new subscriptions and…
— Kimi.ai (@Kimi_Moonshot) July 19, 2026
Discover more
IBL News is funded by the New York-based, family-owned company ibl.ai. Our stories adhere to the highest ethical standards in journalism and are available to news syndication agencies.









