Compare AI models

Specifications
Parameters
671B72B
Context window
128k128k
Benchmarks
SWE-Bench Verified
49.2%60.4%
Overview
CompanyDeepSeekMoonshot AI
Release dateJan 20 2025Jun 17 2025
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Kimi-Dev leads DeepSeek R1 on 1 of the 1 benchmark they both report (SWE-Bench Verified). DeepSeek R1 shipped 148 days before Kimi-Dev, so benchmark comparisons should account for the intervening progress.

DeepSeek R1 has 671B parameters, while Kimi-Dev has 72B. Context windows are 128k (DeepSeek R1) vs 128k (Kimi-Dev).

On SWE-Bench Verified, Kimi-Dev leads at 60.4% vs DeepSeek R1 at 49.2%.