ModelBeat
Models/Kimi K2 Thinking
All models
M

Moonshot: Kimi K2 Thinking

moonshotai/kimi-k2-thinking

Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning.

Context
262K
Input
$0.60 / 1M
Output
$2.50 / 1M
Released
Nov 6, 2025
Modalities
texttext

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Kimi K2 Thinking lands among all models tracked on Model Beat.

24.0
Intelligence Index
Epoch AI
24th percentile of tracked models
29.0
Coding Index
Epoch AI
29th percentile of tracked models
17.0
Agentic Index
Epoch AI
17th percentile of tracked models

Reasoning

3 evals
GPQA Diamond84.2%

Graduate-level scientific reasoning

SimpleQA Verified31.6%

Factual accuracy & hallucination

WeirdML42.8%

Novel ML problem-solving

Coding

1 evals
WebDev Arena1337

Human-rated web development

Math

3 evals
AIME 2024/202583.1%

Olympiad-qualifier math

FrontierMath21.4%

Research-level math problems

FrontierMath Tier 40.0%

Hardest research math

Agentic & Tools

3 evals
APEX4.0%

Multi-step agentic tasks

METR task horizon54 min

Autonomous task length

Terminal-Bench35.7%

Command-line agentic tasks

In the news

4

Model & benchmark data from Epoch AI (CC BY); pricing, specs & descriptions from OpenRouter.