Fourbitfourbit
Request access
LLM

Qwen3 8B

Alibaba · 8B · 256k ctx

Same recipe surface as Llama 8B but with 256k context out of the box. Good fit for retrieval-replacement or long-doc Q&A tunes.

Train this model →Read the recipe
Quickstart

Launch a Qwen3 8B fine-tune

~ fourbit
$ fourbit train \
  --base qwen3-8b \
  --data s3://acme/dataset.jsonl \
  --recipe lora
More LLMs