Kimi models, local and online
Every guide we have written about Moonshot's Kimi family, newest generation first. This page keeps its address when the next version arrives, so it is the one to bookmark.
Every guide
Kimi K3 Explained: Moonshot's 2.8T Open Weight Flagship
Kimi K3 is Moonshot AI's 2.8 trillion parameter MoE with 1M context. Specs, benchmarks, API pricing, and the July 27 open weights date.
Can You Run Kimi K3 Locally? The Honest Hardware Math
Kimi K3 has 2.8 trillion parameters. The honest math on running it locally, what the July 27 weights release changes, and what to run instead.
How to Run Kimi K3: Every Option That Actually Works
How to run Kimi K3 in July 2026: Moonshot API setup with code, OpenRouter in one line, the Kimi app, and the honest math on running the 2.8T model locally.
Kimi K3 online: browser, desktop app and API key
Kimi K3 runs on somebody else's GPUs because nobody has 2.8 trillion parameters at home. Three routes to it, the credit price per token, and the client setup.
Kimi K3 vs Kimi K2.6: What Actually Changed
Kimi K3 (2.8T, 1M context) vs Kimi K2.6 (1T): specs, benchmarks, pricing, and availability compared. Which one you can actually use today, hosted or locally.
Kimi K3 vs DeepSeek V4: price, thinking, vision, plan
One of these reads pictures and one costs almost nothing per token. Where the gap is real, where it is not, and which plans and packs reach all three.
FAQ
Can I run Kimi K3 at home?
No. The smallest usable GGUF is 466 GB on disk, and the biggest consumer machines hold 512 GB of memory. The hardware math is on its own page.
Can I use Kimi K3 without a GPU?
Yes, through a hosted service. Kimi K3 is in the LU Cloud catalog, which opens in a browser.