Announcing our latest research breakthrough: Agent Q - bringing next-generation AI agents with planning and AI self-healing capabilities, with a 340% improvement over LLama 3's baseline zero-shot performance!

In our real-world Agent Q experiments, reaching a success rate of 95.4% in autonomous web agent tasks. Read more about the paper below 👇

To be one of the first to receive access to Agent Q in our developer platform, join our early access waitlist!

Here is the paper

OK, but what I need is something where I can seed it with an initial question, and then let it run for a full week, compounding the context, and then running new searches based on the updated context. So I need it to be autonomous, context aware, and context compounding, and then check back in a week to see what it found.

How are we defining self healing capabilities?

Looking forward to a real-world demo.

bruhhh

Run it on WebArena and report back the number

I can’t find the model card or the data on hugging face, could you give us a link?

@_xjdr is this you