Kimi K3 is listed by Moonshot as a 2.8T-parameter Mixture-of-Experts model with 104B activated parameters per token, 896 experts, 16 selected experts per token, and a 1,048,576-token context length.
The important part is that active parameters do not make the full checkpoint small. For self-hosting, the full weights still need to be stored and served across a large accelerator setup.
AI handoff prompt and permissions
Copy this prompt into Kimi Code, K3Nova, or another AI agent. Keep approval manual for file writes, shell commands, account actions, and secrets.
Copyable AI prompt
Use this to give the agent the task and safety boundary in one message.
You are my AI agent for this task: explain Kimi K3 model size, active parameters, context length, and what those numbers mean for deployment planning.
Start by restating the goal and the permissions you need.
Use official docs or the files I provide before making claims.
Give me a direct answer first, then a short table or checklist.
If commands are needed, show exact copyable commands without a shell prompt.
Ask before writing files, running shell commands, deleting or moving data, logging in, spending money, changing account settings, or handling API keys.
Stop and ask me when a step requires secrets, payment, account access, destructive cleanup, or a permission broader than the task.
Recommended AI permissions
Permission
Give AI
Why
Public research
Allow official docs, cited public pages, and web search results.
Needed to refresh facts without touching private data.
Local files
Do not allow local file access unless you name a folder for inspection.
Most research pages do not need your machine.
Write or shell
Ask before Write, Edit, Bash, or any command execution.
Keeps the task reviewable.
Sensitive actions
Deny login, payment, API keys, account changes, and private screenshots.
Those actions need a human decision.
How to read the facts
Total parameters
2.8T tells you Kimi K3 is frontier-scale and far beyond ordinary single-GPU local models.
Activated parameters
104B active parameters per token helps explain MoE efficiency but does not eliminate the need to store the full expert set.
Context length
1M context supports very long inputs when the product surface and plan allow it, but it should still be prepared carefully.
Quick scan cards
Total parameters
2.8T Full model scale
Activated parameters
104B Approximate active compute per token
Experts
896 total, 16 selected per token Sparse MoE routing
Context length
1,048,576 tokens Long-context ceiling listed in official materials
Quantization
MXFP4 weights and MXFP8 activations Serving format details that affect memory math
FAQ
Is Kimi K3 a 2.8T model?
Yes, official Kimi K3 materials list 2.8T total parameters.
Does 104B active mean it fits like a 104B model?
No. The full weights still need to be stored and served.
Why does context length matter for size?
Long context increases cache and memory pressure during real use.
K3Nova