Websites · PROJECT NOTES
AssistantBench
A live-web research benchmark with realistic information-gathering tasks, public development answers and a SeePlanAct browser-agent implementation.
What can you do with it?
Test whether an agent can plan a web investigation, gather evidence across sites and return a supported answer.
What should you know before using it?
Live websites can change, and test answers are withheld for submission-based evaluation. It is not a resettable website replica. Dataset licensing is Apache 2.0; SeePlanAct code uses OpenRAIL-S with use restrictions. Browser and model access are required.
What is verified here?
This entry was reviewed against original project material. We have not independently run the software or reproduced its results. AI-development claims are author-reported; check upstream for current versions, setup requirements, and licenses.
- Resource type
- Dataset + benchmark
- Provenance
- Development provenance not established
- Reviewed
- Platforms
- Web