DeepSeek-V4.1-FlashvsKimi K2 (0905)

DeepSeek-V4.1-Flash
Kimi K2 (0905)
Specifications
Parameters
1T
Context window
256k
API pricing
Cheapest input
$0.60Novita
Cheapest output
$2.50Novita
Benchmarks
GPQA Diamond
90.9%75.1%
Benchmarks
SWE-Bench Verified
65.8%
DeepSWE 1.1
74.2%
NL2Repo-Bench
65.4%
Terminal-Bench 4.0
31.2%
Terminal-Bench 3.0
30%
Terminal-Bench 2.1
90.6%
CyberGym
88.1%
Humanity's Last Exam · no tools
36.8%
Humanity's Last Exam · with tools
63.9%
Agent's Last Exam · pass@1
31.8%
AutomationBench
54.8%
Chartography · with tools
78.9%
Overview
CompanyDeepSeekMoonshot AI
Release dateSep 10 2026Sep 5 2025
AccessProprietaryOpen Weight

Other comparisons

DeepSeek-V4.1-FlashvsClaude Fable 5.1Kimi K2 (0905)vsClaude Fable 5.1DeepSeek-V4.1-FlashvsGPT-6 AstraKimi K2 (0905)vsGPT-6 AstraDeepSeek-V4.1-FlashvsGemini 3.8 FlashKimi K2 (0905)vsGemini 3.8 FlashDeepSeek-V4.1-FlashvsMuse Spark 1.3Kimi K2 (0905)vsMuse Spark 1.3DeepSeek-V4.1-FlashvsGrok 4.6Kimi K2 (0905)vsGrok 4.6DeepSeek-V4.1-FlashvsMistral Medium 3.5Kimi K2 (0905)vsMistral Medium 3.5

Frequently asked questions

DeepSeek-V4.1-Flash leads Kimi K2 (0905) on 1 of the 1 benchmark they both report (GPQA Diamond). Kimi K2 (0905) shipped 370 days before DeepSeek-V4.1-Flash, so benchmark comparisons should account for the intervening progress.

DeepSeek-V4.1-Flash is proprietary, while Kimi K2 (0905) is open weight.

On GPQA Diamond, DeepSeek-V4.1-Flash leads at 90.9% vs Kimi K2 (0905) at 75.1%.