Comparing 2 models

Claude Sonnet 5 vs DeepSeek V4 Flash

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

metric
Model 1Claude Sonnet 5
Model 2DeepSeek V4 Flash
Best for
Autonomous Agents (#4)
Code Completion (#2)
Provider
anthropic
deepseek
Context
1M
1.0MBest
Input / 1M
$2
$0.098Best
Output / 1M
$10
$0.20Best
Tool use
yesBest
yesBest
Vision
yesBest
Reasoning
yesBest
yesBest
Open-weight
Released
Jun 30, 2026
Apr 24, 2026