# UI-R1 3B

A Qwen2.5-VL checkpoint trained with reinforcement learning for GUI action prediction and evaluated on mobile, desktop and web grounding tasks.

- Catalog URL: https://awesomecua.github.io/resources/url-f13474d7aa734a30624d/
- Original source: https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1
- Kind: model
- Review: curated
- Scope: core
- Environments: web, desktop, android
- Topics: grounding, reinforcement-learning, action-prediction
- Evidence checked: 2026-10-17T02:03:33Z
- ID: url:f13474d7aa734a30624d

## Limitations

Two safetensors shards and tokenizer/configuration files are publicly ungated. The model card declares MIT; upstream base-model terms are separate. The mutable card compares v1, v2 and UI-R1-E results without pinning every result to a commit, so this record identifies the Hub release and does not import an unversioned score.

- Record JSON: https://awesomecua.github.io/resources/url-f13474d7aa734a30624d/record.json

## License declarations

- code: not recorded
- data: not recorded
- document: not recorded
- weights: mit — https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md

## Primary evidence

- https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md (retrieved 2026-10-17T02:03:33Z; SHA-256 620fbee20a6002bbc13562030ce4988d41ef17b0c033ec6872d83741f597adb8)
- https://huggingface.co/api/models/LZXzju/Qwen2.5-VL-3B-UI-R1 (retrieved 2026-10-17T02:03:33Z; SHA-256 6a206702a1a4faef428460e07f52aaf8fd1f1a144d711d807ca42479b248f697)
- https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md (retrieved 2026-10-16T06:36:00Z; SHA-256 f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3)
- https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md (retrieved 2026-10-17T02:03:33Z; SHA-256 620fbee20a6002bbc13562030ce4988d41ef17b0c033ec6872d83741f597adb8)