Comparing 2 models

DeepSeek V4 Flash vs Kimi K3

Facts first, then task-specific rankings. A highlighted cell is better only for that row, not universally.

metric
Model 1DeepSeek V4 Flash
Model 2Kimi K3
Best for
Code Completion (#2)
Code Completion (#1)
Provider
deepseek
moonshotai
Context
1.0MBest
1.0MBest
Input / 1M
$0.098Best
$3
Output / 1M
$0.20Best
$15
Tool use
yesBest
yesBest
Vision
yesBest
Reasoning
yesBest
yesBest
Open-weight
Released
Apr 24, 2026
Jul 16, 2026