DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on...
Rank
#1
Down 31% week over week
Tokens
25.7T
2026-10-21
Requests
643.5M
643,500,422
Tool calls
-
Input / M
$0.30
Output / M
$1.20
Cache / M
$0.006
Context
1.0M
1,048,576 tokens