Hugging Face Clones OpenAI's Deep Research In 24 Hours 17
An anonymous reader quotes a report from Ars Technica: On Tuesday, Hugging Face researchers released an open source AI research agent called "Open Deep Research," created by an in-house team as a challenge 24 hours after the launch of OpenAI's Deep Research feature, which can autonomously browse the web and create research reports. The project seeks to match Deep Research's performance while making the technology freely available to developers. "While powerful LLMs are now freely available in open-source, OpenAI didn't disclose much about the agentic framework underlying Deep Research," writes Hugging Face on its announcement page. "So we decided to embark on a 24-hour mission to reproduce their results and open-source the needed framework along the way!"
Similar to both OpenAI's Deep Research and Google's implementation of its own "Deep Research" using Gemini (first introduced in December -- before OpenAI), Hugging Face's solution adds an "agent" framework to an existing AI model to allow it to perform multi-step tasks, such as collecting information and building the report as it goes along that it presents to the user at the end. The open source clone is already racking up comparable benchmark results. After only a day's work, Hugging Face's Open Deep Research has reached 55.15 percent accuracy on the General AI Assistants (GAIA) benchmark, which tests an AI model's ability to gather and synthesize information from multiple sources. OpenAI's Deep Research scored 67.36 percent accuracy on the same benchmark with a single-pass response (OpenAI's score went up to 72.57 percent when 64 responses were combined using a consensus mechanism).
As Hugging Face points out in its post, GAIA includes complex multi-step questions such as this one: "Which of the fruits shown in the 2008 painting 'Embroidery from Uzbekistan' were served as part of the October 1949 breakfast menu for the ocean liner that was later used as a floating prop for the film 'The Last Voyage'? Give the items as a comma-separated list, ordering them in clockwise order based on their arrangement in the painting starting from the 12 o'clock position. Use the plural form of each fruit." To correctly answer that type of question, the AI agent must seek out multiple disparate sources and assemble them into a coherent answer. Many of the questions in GAIA represent no easy task, even for a human, so they test agentic AI's mettle quite well. Open Deep Research "builds on OpenAI's large language models (such as GPT-4o) or simulated reasoning models (such as o1 and o3-mini) through an API," notes Ars. "But it can also be adapted to open-weights AI models. The novel part here is the agentic structure that holds it all together and allows an AI language model to autonomously complete a research task."
The code has been made public on GitHub.
Similar to both OpenAI's Deep Research and Google's implementation of its own "Deep Research" using Gemini (first introduced in December -- before OpenAI), Hugging Face's solution adds an "agent" framework to an existing AI model to allow it to perform multi-step tasks, such as collecting information and building the report as it goes along that it presents to the user at the end. The open source clone is already racking up comparable benchmark results. After only a day's work, Hugging Face's Open Deep Research has reached 55.15 percent accuracy on the General AI Assistants (GAIA) benchmark, which tests an AI model's ability to gather and synthesize information from multiple sources. OpenAI's Deep Research scored 67.36 percent accuracy on the same benchmark with a single-pass response (OpenAI's score went up to 72.57 percent when 64 responses were combined using a consensus mechanism).
As Hugging Face points out in its post, GAIA includes complex multi-step questions such as this one: "Which of the fruits shown in the 2008 painting 'Embroidery from Uzbekistan' were served as part of the October 1949 breakfast menu for the ocean liner that was later used as a floating prop for the film 'The Last Voyage'? Give the items as a comma-separated list, ordering them in clockwise order based on their arrangement in the painting starting from the 12 o'clock position. Use the plural form of each fruit." To correctly answer that type of question, the AI agent must seek out multiple disparate sources and assemble them into a coherent answer. Many of the questions in GAIA represent no easy task, even for a human, so they test agentic AI's mettle quite well. Open Deep Research "builds on OpenAI's large language models (such as GPT-4o) or simulated reasoning models (such as o1 and o3-mini) through an API," notes Ars. "But it can also be adapted to open-weights AI models. The novel part here is the agentic structure that holds it all together and allows an AI language model to autonomously complete a research task."
The code has been made public on GitHub.
Wont anyone think of Sam Altman? (Score:3)
Re: Wont anyone think of Sam Altman? (Score:2)
What's going on with the investigation on that Suchir fellow?
Can't say I'm upset about this (Score:2, Interesting)
wow, glad I sat this round out (Score:3)
I'm announcing my new AI model: Deeper Open Face Hugging Deeper Seeker.
It's more open and goes deeper than any AI has gone yet.
Seriously, It's an AI model, not a porn movie. For real.
Re: (Score:1)
Too bad, because there is probably a huge demand for an AI that can do decent porn movies on your home computer with modest graphic card.
Re: (Score:2)
South Korean middle schoolers have been slapped down for something nearly like that. Using generative AI to face swap classmates into porn movies.. I think middle schoolers everywhere are working on it
Seriously, apologies to the girls who don't deserve this, though, but you can't stop water from running down a hill.
Don't yell at me, I didn't do it.
Re: (Score:1)
Were they exclusively swapping faces of female classmates? No male face swapping at all?
Interesting datapoint, if so.
Face Hugger Latched Onto OpenAI (Score:3)
Here's hoping that we don't see any Xenomorphs bursting forth.
Re: (Score:2)
I'm more terrified of the clones than the face huggers. Clones are scary.
Sorta (Score:2)
But it is true and increasingly clear that you can stay "pretty close" to the frontier with orders of magnitude fewer resources than it took to blaze the trail in the first place.
If
It's so bad to create something and have it taken (Score:1)
Isn't it OpenAI?
So is it worse to have OpenAI take my words and artwork to use as their own?
Or for Someone else to take OpenAI's work and make it available to everyone.
I'm not feeling much sympathy for OpenAI.
Re: (Score:2)
I'm not sure you're making the point you think you are. If you make something and someone else says "that's a neat idea, I'm going to make one too" that's a good thing. If you object, you're an entitled dick.
Re: (Score:1)
My point is OpenAI ignored robots.txt and violated websites terms of service and scraped the entire internet taking many people's copyrighted guides, how-to's, stories.
And now OpenAI is complaining that Deepseek basically did the same thing to OpenAI (using OpenAI to train Deepseek)...and OpenAI is complaining that Deepseek violated their terms of service.
Which is hypocritical.
And if given a choice between a hypocritical actor who violated terms of service and common norms or an actor who violated terms of
Re: (Score:2)
This is a story about a bunch of people on Hugging Face making something that's similar to something OpenAI has described. There are no terms of service violations, alledged or otherwise, and DeepSeek isn't involved.
Maybe you're posting on the wrong story?