Kimi K3

moonshotai/kimi-k3

Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work

Compare

Context window: 1M
Input: Image · Text · Video
Output: Text
Released: 2026-07-16

Max output: 131K
Model parameters: 2.8T
Weights: Open
License: kimi-k3
Capabilities: Attachments, Open weight, Reasoning, Structured, Tool call

Pricing

No official pricing available.

Estimate usage cost

Compare verified public prices for this exact model and region. Enter uncached input and cache usage separately. Estimates exclude unspecified fees, credits and currency conversion. Set requests to zero to estimate token charges only.

Input tokens

Output tokens

Cache read tokens

Cache write tokens

Requests

Context tokens (optional)

Currency

Batch processing

Performance

Overall capability

55.0/100

Coding: 48.5
Reasoning: 54.1
Professional work: 55.9
Knowledge & language: 58.8


Arena Agent Overall	Professional work	LMArena	2026-11-08
APEX-Agents 1.1	Professional work	Mercor	2026-11-05
Chess Puzzles	Reasoning	Epoch AI	2026-11-05
Terminal-Bench 4.0 (AA rerun)Kimi K3 (Low)	Coding	Artificial Analysis	2026-11-05
AA-OmniscienceKimi K3 (Low)	Knowledge & language	Artificial Analysis	2026-11-05
MCP Atlasmax	Professional work	Scale AI	2026-11-05
Terminal-Bench Science · AAKimi K3 (Max)	Reasoning	Artificial Analysis	2026-11-05
Finance Agent	Professional work	Vals AI	2026-11-05

Providers

No providers found

ModelRegistry · Catalog updated 2026-11-10