tev1-4B-experimental

Back to Devoured - September 24, 2026

DEVOURED

tev1-4B-experimental

Together AI released tev1-4B-experimental, a Jev-like classifier fine-tuned on Qwen3.5 4B that cost only $17 to train.

What: The model is available on Together's serverless platform for $0.042 per million input tokens. Together AI also published the data recipe and a tutorial for fine-tuning custom classifiers.

Why it matters: This highlights the increasing accessibility of training specialized, small-parameter models for specific classification tasks at a negligible cost.

Takeaway: If you need a low-cost, dedicated classifier, follow Together AI's guide to fine-tune your own version using their released data recipe.

Decoder

Jev-like: Refers to a specific classifier architecture or evaluation style focused on classification precision.

Original article

We're releasing tev1-4B-experimental, a Jev-like classifier finetuned on top of Qwen3.5 4B.

We're making it available on Together serverless at $0.042/M input & $0/M output.

Also releasing the data recipe & a tutorial on how to finetune your own (Tev1 cost $17 to train!).

https://twitter.com/nutlope/status/2102881280115249597

Note that this is an experimental model meant to show how you can train your own specialized decision models fairly easily!

You can also check out the model weights here, along with the full tutorial on finetuning your own classifiers above: huggingface.co/togethercomputer/Tev1-4B-experimental

Explore all 53 articles in today's digest