/ Models

Language models

MoonshotAI: Kimi K3

Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...

Maker
Moonshotai
Context window
1M
Max output
Input / 1M
$3
Output / 1M
$15
1M in + 1M out
$18

Everything else

Max output
Input / 1M$3
Output / 1M$15
Cache read / 1M$0.3
Cache write / 1M
1M in + 1M out$18
Takestext, image
Returnstext
Toolsyes
Reasoningoptional · max, high, low
Open weightsyes
Released2026-07-16
Knowledge cutoff
Retires
Catalog idmoonshotai/kimi-k3

Benchmarks

sourceepoch
eci155.59
confidenceSpeculative
accessibilityOpen weights (non-commercial)
gpqa0.931
gpqaStderr0.0149
aime0.972
aimeStderr0.0112
simpleQA0.427
simpleQAStderr0.0156

Benchmark scores by Epoch AI, CC BY 4.0.

Using it

In 00, this model is picked per agent — and per tier, so the model that answers your customers need not be the one that writes your code. Get 00.

← all models