Compare AI models

Specifications
Parameters
671B1T
Context window
128k256k
API pricing
Cheapest input
—$0.49AtlasCloud
Cheapest output
—$2.50AtlasCloud
—$223AtlasCloud · int4
Benchmarks
SWE-Bench Verified
49.2%76.8%
LiveCodeBench
65.9%85%
GPQA Diamond
71.5%87.6%
Overview
CompanyDeepSeekMoonshot AI
Release dateJan 20 2025Jan 27 2026
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Kimi K2.5 leads DeepSeek R1 on 3 of the 3 benchmarks they both report (SWE-Bench Verified, LiveCodeBench, GPQA Diamond). DeepSeek R1 shipped 372 days before Kimi K2.5, so benchmark comparisons should account for the intervening progress.

DeepSeek R1 has 671B parameters, while Kimi K2.5 has 1T. Context windows are 128k (DeepSeek R1) vs 256k (Kimi K2.5).

On SWE-Bench Verified, Kimi K2.5 leads at 76.8% vs DeepSeek R1 at 49.2%. On LiveCodeBench, Kimi K2.5 leads at 85% vs DeepSeek R1 at 65.9%. On GPQA Diamond, Kimi K2.5 leads at 87.6% vs DeepSeek R1 at 71.5%.