modelbenchmark.io

Llama 3.3 70B Instruct abliterated

Meta · llama-3-3-70b-instruct-abliterated

Compare

Open Llama instruction model for multilingual chat, reasoning, and coding

Specification

most-agreed values

Context
33K
Max output
16K
Released
2025-08-08
Knowledge cutoff
Retires
Open weights
yes
Input
text
Output
text

Price

US dollars per million tokens · most-agreed

Input
$0.70
Output
$0.70
Cache read
$0.35

Available from 1 host

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
nano-gpt · huihui-ai$0.70$0.70$0.3533K16K

More from Meta

most-hosted first