# WebRL outcome reward model 8B

The outcome-supervision checkpoint used by WebRL to score browser trajectories during policy training.

- Catalog URL: https://awesomecua.github.io/resources/url-d1e85aeb6d96859c109d/
- Original source: https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b
- Kind: model
- Review: curated
- Scope: core
- Environments: web
- Topics: reward-models, reinforcement-learning, web-navigation
- Evidence checked: 2026-10-17T02:05:40Z
- ID: url:d1e85aeb6d96859c109d

## Limitations

The official API resolves the former THUDM namespace to zai-org and lists four safetensors shards. No model card or weight-license declaration is present in the inspected manifest. The official WebRL README identifies this as its outcome-supervised reward model; it is a training component rather than an action policy.

- Record JSON: https://awesomecua.github.io/resources/url-d1e85aeb6d96859c109d/record.json

## License declarations

- code: not recorded
- data: not recorded
- document: not recorded
- weights: not recorded

## Primary evidence

- https://huggingface.co/api/models/zai-org/webrl-orm-llama-3.1-8b (retrieved 2026-10-17T02:05:40Z; SHA-256 8bafcdf63fe484b6005ca7082da1310459f8c87d4c011e84293a3cc8fca6ae9e)
- https://huggingface.co/api/models/THUDM/webrl-orm-llama-3.1-8b (retrieved 2026-10-17T02:02:29Z; SHA-256 8bafcdf63fe484b6005ca7082da1310459f8c87d4c011e84293a3cc8fca6ae9e)
- https://raw.githubusercontent.com/THUDM/WebRL/HEAD/README.md (retrieved 2026-10-17T02:02:06Z; SHA-256 eb217b96c54c3c2ccfd3a27255c6fd740df40c2d6743ae811858339bf7f740b0)
- https://huggingface.co/api/models/zai-org/webrl-orm-llama-3.1-8b (retrieved 2026-10-17T02:05:40Z; SHA-256 8bafcdf63fe484b6005ca7082da1310459f8c87d4c011e84293a3cc8fca6ae9e)