Introducing Kimi K3 has been revealed and it is open Frontier Intelligence
🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost
🔹 Built for long-horizon agentic coding and self-evolving workflows
Kimi K3 is now live on on http://Kimi.com, Kimi Work, Kimi Code, and the Kimi API.
Open Weights by July 27, 2026.
K3 is built on Kimi Delta Attention (KDA) and Attention Residuals (AttnRes), two architectural updates designed to improve how information flows across sequence length and model depth.
They have also scaled up Mixture of Experts (MoE) sparsity, effectively activating 16 out of 896 experts when paired with a Stable LatentMoE framework.
Together with refined training and data recipes, these structural changes yield an approximate 2.5× improvement in overall scaling efficiency compared to K2, allowing the model to convert compute into intelligence more effectively.
Introducing Kimi K3: Open Frontier Intelligence
🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal
🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts
🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional… pic.twitter.com/eFHEbdxn3P— Kimi.ai (@Kimi_Moonshot) July 16, 2026
K3 is built on Kimi Delta Attention (KDA) and Attention Residuals (AttnRes), two architectural updates designed to improve how information flows across sequence length and model depth.
We have also scaled up Mixture of Experts (MoE) sparsity, effectively activating 16 out of… pic.twitter.com/yyFMLACiek
— Kimi.ai (@Kimi_Moonshot) July 16, 2026
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
My initial thoughts, SpaceX/XI is taking the smart route, focusing on hardware layer. They can then run any model based on demand. Anthropic is toast, OpenAI is more flexible than Anthropic business-wise but still not looking good for them. Google, Amazon and Microsoft can’t be counted out yet as they have a lot of hardware and existing datacenters that can migrate from storage to a combo of storage + AI hardware. Ultimately though SpaceX is going to win in the West.
China on the other hand has enourmous power production capability and growing … so only a matter of time until they are neck and neck with hardware, then they will undercut everyone, except perhaps SpaceX who will not be constrained in terms of power or datacenter space.
Expect OpenAI and Anthropic to merge with Google, Amazon, or Microsoft when the time is right – probably post-IPO.
As Chinese models approach performance of frontier models, their API cost too. The real risk for US AI labs now is big customers will switch to open source models because they can be customized and it’s much cheaper to run models locally. Even retail market might not safe for US AI labs because one small company can still host an open source model then sell tokens on retail market at cheaper price.
So SpaceX will be worth more than Earth…
But Kimi K3 just leapfrogged Grok 4.5.