Qwen3-1.7B-Wordle-SFT
Prime fine-tuned Qwen3-1.7B for the Wordle environment. This supervised checkpoint teaches the interaction format before multi-turn reinforcement learning.
Research release
What the release provides
The release is part of the prime-rl Wordle example. Its immediate role is to teach the interaction format before training a stronger game policy.Qwen3-1.7B-Wordle-SFT official model cardPrime RL Wordle training example
How to use it well
Practical guidance. Check whether the model submits legal guesses and reads the feedback. Format accuracy and game success are different outcomes.
Game traces teach the format
Prime's example uses willcb/V3-wordle, a dataset of game traces. It publishes the supervised model after a short warmup, before the separate reinforcement-learning run.Prime RL Wordle training example
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Repository license | Apache 2.0Qwen3-1.7B-Wordle-SFT official model card |
Related models
- Qwen3-1.7B-Wordle-RL
Prime Intellect ยท Prime RL research models
A Qwen3 model trained to improve Wordle play through multi-turn rewards.
Launch date not confirmedResearch releaseDownloadable