Kimi K3

Kimi K3 is Moonshot AI's open-weight, always-thinking multimodal model for long-context reasoning, coding, tool use, and native image/video understanding.

Reasoning
Vision
Video Input
Tool Calling
Structured Output

Added Jul 16, 2026

Model weights

Pricing

Auto routing · per 1M tokens

Input: $1.80
Output: $9.00
Cache read: $0.18

Compare provider prices

Specifications

Context window: 1M
Max output: 1M
Avg output (7d): 1.1K tokens; Longer than 74% of models

Benchmarks

Artificial Analysis

Design Arena

Sourced from Artificial Analysis.

Intelligence Index

43.6

Better than 94% of models compared

Coding Index

76.2

Better than 93% of models compared

Agentic Index

50.0

Better than 90% of models compared

Agentic work

AutomationBench-AA

Workflow automation with guardrail penalties

58.3%

Better than 82% of models compared

AutomationBench-AA Tasks Completed

Fully completed workflows without guardrail violations

24.8%

Better than 30% of models compared

Harvey LAB-AA

Legal agentic work criterion pass rate

94.6%

Better than 96% of models compared

AA-Briefcase

Agentic knowledge work (Elo)

1501 Elo

Better than 86% of models compared

GDPval-AA v2

Economically valuable tasks (Elo)

1538 Elo

Better than 85% of models compared

Document reasoning

GDP.pdf

Professional PDF reasoning: all-pass rate

22.0%

Better than 76% of models compared

AA-LCR v1.1

Long context reasoning with updated grading

88.7%

Better than 99% of models compared

MLCR-AA

Medical long-context reasoning

38.3%

Better than 84% of models compared

Reasoning

HLE

Humanity's Last Exam

46.9%

Better than 94% of models compared

CritPt

Research-level physics reasoning

23.4%

Coding

Terminal-Bench v4.0

Practical coding and terminal tasks

12.6%

Better than 66% of models compared

SciCode

Python programming for scientific computing

59.5%

Better than 95% of models compared

Knowledge

AA-Omniscience Accuracy

Proportion of correctly answered questions

47.6%

AA-Omniscience Hallucination Rate

Rate of incorrect answers among non-correct responses

53.2%

Legacy benchmarks

GPQA Diamond (legacy)

Graduate-level scientific reasoning

93.5%

Better than 97% of models compared

AA-LCR (unversioned / legacy)

Long context reasoning evaluation

88.7%

Better than 99% of models compared

GDPval-AA (unversioned / legacy)

Economically valuable tasks

51.9%

Last updated Oct 5, 2026