Assertion Supported AI assessment confidence: 95% certainty 4/5 debate potential 2/5

Fulford: Deep Research Completes Multi-Hour Human Work in 5 to 30 Minutes

Isa Fulford · No Priors Ep. 112 | With OpenAI Deep Research, Isa Fulford · Apr 24, 2025 · at 27:01

Isa Fulford of OpenAI explains the performance efficiency of Deep Research to Sarah Guo.

0:00 / 0:07exact quote · 7.2s

▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark

“Right now, in five or 30 minutes, it can do what human experts rate take many hours.”

quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →

The claim, plainly

In 5 to 30 minutes, OpenAI Deep Research can complete tasks that human experts evaluate as taking many hours.

a quote alone can be ambiguous, so this restates it with the references and context of the conversation around it resolved. AI-written; the tape below the player is the source.

Assessment

OpenAI's official technical release materials and system card for Deep Research (published in February 2025) document both elements of the claim: the system operates within a runtime window of 5 to 30 minutes, and technical evaluations explicitly measured performance against tasks rated by human experts as requiring multiple hours (categorized in tiers such as 1–3 hours, 4–6 hours, and 7–9 hours of expert effort). External benchmarking on datasets like GAIA and evaluations by independent third parties confirmed the agentic runtime and completion of multi-hour research workflows.

OpenAI's February 2025 'Introducing Deep Research' launch documentation specifies an autonomous runtime window of 5 to 30 minutes.
OpenAI's technical evaluations categorized benchmark tasks under 'Pass Rate on Expert-Level Tasks by Estimated Hours' spanning 1–3, 4–6, and 7–9 hours of estimated human expert effort.
The Deep Research System Card (February 25, 2025) and independent benchmark evaluations (e.g., GAIA, FutureSearch) confirmed the system executes multi-step expert research tasks within this operational runtime.

Checked against

[1] businessinsider.com [2] openai.com [3] openai.com [4] openai.com [5] openai.com [6] centerconsulting.com [7] scribd.com [8] openai.com

OpenAI Deep Research system card 2025 "deep research" "5 to 30 minutes" site:openai.com OpenAI Deep Research February 2025 announcement runtime 5 to 30 minutes hours "deep research" "tens of minutes" site:openai.com "Pass Rate on Expert-Level Tasks" OpenAI "deep research" "many hours" site:openai.com OpenAI Deep Research technical report system card GAIA benchmark runtime minutes hours "Deep Research" "GAIA" "67.36" "Introducing Deep Research" "GAIA" "Humanity's Last Exam" site:openai.com OpenAI Deep Research GAIA benchmark human expert hours

assessed as of 2026-09-03

This assessment is an AI reading of a public statement, made against the cited references with the confidence shown. It can be wrong. The quote, the tape and the references are all on this page; if something is off, use report an error below. full disclaimer →

Mentioned

Deep Research human experts

⚑ Spot an error? Report it

More from Isa Fulford

Insight

Fulford: Information Synthesis Is a Prerequisite for Scientific AI Discovery

“Secondly, I think the overall goal for OpenAI is to create an AGI that can make new scientific discoveries, and we kind of felt that a prerequisite to that is to be able to synthesize information. You know, if you can't write a literature review, you're not go…”

Isa Fulford Apr 24, 2025 ▶ 2:58 No Priors Ep. 112 | With OpenAI Deep Research, Isa Fulford

Insight

Fulford: RFT is only worthwhile for out-of-distribution or make-or-break tasks

“I think if you have a very specific task that you think is so different to anything that the model was likely trained on and you try it a bunch of times yourself and you've tried a lot of different prompts and it's just really not good at it. So maybe it's gen…”

Isa Fulford Apr 24, 2025 ▶ 7:50 No Priors Ep. 112 | With OpenAI Deep Research, Isa Fulford

Assertion Not checkable as stated

OpenAI: Deep Research learned upfront planning without explicit instruction

“We didn't teach it To plan up front, but sometimes we'll see it does end up making a plan up front before starting its research.”

Isa Fulford Apr 24, 2025 ▶ 10:30 No Priors Ep. 112 | With OpenAI Deep Research, Isa Fulford

Assertion Not checkable as stated

Fulford: OpenAI Deep Research attempts reward hacking around tool restrictions

“Sometimes the model will do smart things and try to get around restrictions you put on it. So you have to make sure that it's not hacking, you know, and trying to use a different search engine other than the search engine that you gave it or something like tha…”

Isa Fulford Apr 24, 2025 ▶ 10:42 No Priors Ep. 112 | With OpenAI Deep Research, Isa Fulford

Assertion Not checkable as stated

Fulford: Deep Research Hallucinates Less Than Any Prior OpenAI Model

“While this model is hallucinates less than any model that we've ever released, it is still possible for it to hallucinate most times because it will infer something incorrectly from one of its sources.”

Isa Fulford Apr 24, 2025 ▶ 11:38 No Priors Ep. 112 | With OpenAI Deep Research, Isa Fulford

Disclosure

Fulford: OpenAI Prioritized Read-Only Synthesis Over Action Agents

“Yeah, so I think before we focused on taking right actions, which those are examples of taking right actions, we wanted to get really good at synthesizing information from a large number of sources and mostly read-only tasks.”

Isa Fulford Apr 24, 2025 ▶ 2:38 No Priors Ep. 112 | With OpenAI Deep Research, Isa Fulford

Fulford: Deep Research Completes Multi-Hour Human Work in 5 to 30 Minutes

More from Isa Fulford

Turn any episode into a week of clips.