#55
Kimi K2 Thinking
kimi-k2-thinking
0.0
Helpfulness
Instruction Following
Comprehension
Empathy
Creative Writing
Helpfulness
0.0
Instruction Following
0.0
Comprehension
0.0
Empathy
0.0
Creative Writing
0.0
Speed
Avg 42.5 tok/s
Release Date
November 6, 2025
Lab
Moonshot AI
Type
Open Source
Context Size
128K
Max Output Tokens
4.1K
Cost per 1 million tokens
$0.40 / $2.00
Model Inputs*
Text
Model Outputs*
Text
Tool Calling*
Enabled
Overall Assistant Score
An average score combining the 5 main categories.
78.44 pts
Rank #55
21th Percentile
0
Novice
33
Capable
66
Proficient
100
Expert
Kimi K2 Thinking is Moonshot AI's dedicated reasoning model, designed as a 'thinking agent' that processes complex queries step-by-step. With a 1-trillion parameter MoE architecture (activating 32B parameters), it delivers high instruction following (88.5) and comprehension (89.5). While its creative output (76.2) and empathy (70.5) are secondary to its logic-first design, it offers a powerful open-source alternative for high-precision tasks. Its specialized reasoning process results in slower speeds (42.5) but ensures depth in logical deduction, outperforming many proprietary models in instruction adherence at a significantly lower cost.

Overall Score vs Price

Price axis is logarithmic (USD per 1M input tokens) · Higher score and lower price is better

Ringed point is this model

Anthropic
DeepSeek
Google
Meta
Moonshot
OpenAI
SpaceX AI
Zhipu AI

Intelligence

Overall Score · Higher is better

Speed

Output tokens per second · Higher is better

Price

USD per 1M input tokens · Lower is better

Moonshot Models

Overall Score · Same provider

Moonshot Family

Overall Score · Same model family

Closest Rivals

Overall Score · Nearest by overall rank