# InfiGUI-G1 7B

A GUI-grounding checkpoint trained with adaptive exploration rewards and multiple candidate answers to improve instruction-to-element alignment.

- Catalog URL: https://awesomecua.github.io/resources/url-d70a755e3b2102f25f34/
- Original source: https://huggingface.co/InfiX-ai/InfiGUI-G1-7B
- Kind: model
- Review: curated
- Scope: core
- Environments: cross-platform
- Topics: grounding, reinforcement-learning
- Evidence checked: 2026-10-17T02:44:29Z
- ID: url:d70a755e3b2102f25f34

## Limitations

Four safetensors shards are public. The model card declares Apache-2.0 and identifies Qwen2.5-VL-7B-Instruct as its base. It is the G1 grounding checkpoint, distinct from the previously cataloged InfiGUI-R1 reasoning checkpoint; benchmark scores remain publisher claims.

- Record JSON: https://awesomecua.github.io/resources/url-d70a755e3b2102f25f34/record.json

## License declarations

- code: not recorded
- data: not recorded
- document: not recorded
- weights: apache-2.0 — https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md

## Primary evidence

- https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md (retrieved 2026-10-17T02:44:29Z; SHA-256 253055f0f77266a64d2db5de8fd04c4202f4cda3be733dd1b00d8565116dd240)
- https://huggingface.co/api/models/InfiX-ai/InfiGUI-G1-7B (retrieved 2026-10-17T02:44:29Z; SHA-256 10278eb0f1600557e432a0f8dde2c40cfffe12330ddb949b6a0fc02bfbc6c9ee)
- https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md (retrieved 2026-10-16T06:36:00Z; SHA-256 4b68c82a2dd0063a450bde34c8bffbd931565a91a2982a5b47be7e711815b284)
- https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md (retrieved 2026-10-17T02:44:29Z; SHA-256 253055f0f77266a64d2db5de8fd04c4202f4cda3be733dd1b00d8565116dd240)