About
What it is for
A 36.7M-parameter model trained from scratch, mostly on English. It runs in your browser.
Wren 40 is our own small language model: about 36.7 million parameters, pre-trained from scratch on 2 billion tokens of mostly English text, then tuned on 400 million tokens of conversation data.
It runs in your browser at ai.avenpa.com, so the first visit downloads the model and nothing you type leaves your device. It answers short English questions and simple code. Korean, arithmetic and fact checks are still weak, and it can be wrong.
In short
Wren 40, briefly
01
From scratch
Weights, tokenizer and training loop are ours.
02
In your browser
Runs on your device; nothing you type is sent to a server.
03
36.7M parameters
2B tokens of pre-training, 400M of chat tuning.
Questions about Wren 40?
Write to the studio. A person who works on it will answer.
Contact the studio