SkyworkAI / agent-studio

Benchmarks, environments, and toolkits for general computer agents
https://skyworkai.github.io/agent-studio/
GNU Affero General Public License v3.0
153 stars 11 forks source link

Would love to see conversational web agents #37

Closed xhluca closed 5 months ago

xhluca commented 5 months ago

Would love to see conversational web navigation benchmarks like WebLINX and MT-Mind2Web!

ltzheng commented 5 months ago

Thanks for bringing this up! Conversational web agents are an exciting field, and WebLINX and MT-Mind2Web are indeed important contributions. We'll add these benchmarks in our paper's next revision.

xhluca commented 5 months ago

Looking forward it!