Kimi-K2.6

Kimi-K2.6

moonshotai/Kimi-K2.6

About Kimi-K2.6

Kimi K2.6 is an open-source, native multimodal agentic model by Moonshot AI, achieving open-source state-of-the-art on benchmarks including HLE with tools, SWE-Bench Pro, and BrowseComp. Built on a MoE architecture with 1T total parameters and 32B activated, the model supports a 256K-token context window and multimodal inputs (image and video) via its MoonViT vision encoder. K2.6 is optimized for agentic workloads: it sustains 4,000+ tool calls over 12+ hours of continuous execution, scales to 300 parallel sub-agents × 4,000 steps per run to produce 100+ files from a single prompt, and supports both Thinking and Instant inference modes with function calling and multi-turn Preserve Thinking

Available Serverless

Run queries immediately, pay only for usage

Input Price

$

0.95

/ M Tokens

Cache Read

$

0.16

/ M Tokens

Output Price

$

4.0

/ M Tokens

Metadata

Create on

License

MODIFIED MIT

Provider

Moonshot AI

HuggingFace

Specification

State

Available

Architecture

Mixture-of-Experts

Calibrated

No

Mixture of Experts

Yes

Total Parameters

1000B

Activated Parameters

32B

Reasoning

No

Precision

FP8

Context length

262K

Max Tokens

262K

Supported Functionality

Serverless

Supported

Serverless LoRA

Not supported

Fine-tuning

Not supported

Embeddings

Not supported

Rerankers

Not supported

Support image input

Supported

JSON Mode

Supported

Structured Outputs

Not supported

Tools

Supported

Fim Completion

Not supported

Chat Prefix Completion

Supported

Ready to accelerate your AI development?

Ready to accelerate your AI development?

Ready to accelerate your AI development?