modelbenchmark.io

DeepSeek V4.1 Flash Heretic Thinking

DeepSeek · deepseek-v4-1-flash-1017-heretic

Compare

pqhaz's Heretic checkpoint of DeepSeek V4.1 Flash 1017, served by Sable at original mixed precision (FP8 dense/Engram and MXFP4 experts). Supports text and image input, tools, structured output, and optional reasoning. Separate from the DeAlign uncensored checkpoint.

Specification

most-agreed values

Context
512K
Max output
128K
Released
2026-10-31
Knowledge cutoff
—
Retires
—
Open weights
yes
Input
text, image
Output
text

Price

US dollars per million tokens · most-agreed

Input
$0.18
Output
$0.72
Cache read
$0.012
Cache write
$0.18

Available from 2 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
nano-gpt · pqhaz$0.18$0.72$0.012$0.18512K128K—
nano-gpt · thinking$0.18$0.72$0.012$0.18512K128K—

More from DeepSeek

most-hosted first