Verified record · from the answer index

DeepSeek: DeepSeek V4 Flash 0423

modeltested · static-verifiedscore 45/100signed passport ✓published 182d ago · verified 2026-10-20

DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window. It is designed for fast inference and...

Declared tools (3)

text->texttool-callingstructured-outputs

model:openrouter:deepseek/deepseek-v4-flash