# OpenAI Deep Research

13 Hacker News commenters recommended OpenAI Deep Research and 15 criticized it, November 2025 to November 2026. It was recommended in 8 of 13 months.

**Verdict:** People recommend it for thorough, flexible research that surfaces far more sources than a search engine. The most common caveat is that Pro subscribers get only five full runs per month, after which critics say the limited version is weak.

## Why people recommend it

- Strong overall quality and accuracy (4 comments): "Very few mistakes and lots more flexibility and understanding than the travel agents I used before." ([source](https://news.ycombinator.com/item?id=48458177)) "the OpenAI offering has become genuinely amazing" ([source](https://news.ycombinator.com/item?id=46125159)) "the recommendations still seem mostly “correct”" ([source](https://news.ycombinator.com/item?id=45795012))
- Broad, fast source surveying (2 comments): "As it can quickly survey hundreds of sources for overall relevance" ([source](https://news.ycombinator.com/item?id=47711706)) "It is absolutely a better-Google-than-Google class of resource." ([source](https://news.ycombinator.com/item?id=47140842))
- Better than other research tools (2 comments): "It is absolutely a better-Google-than-Google class of resource." ([source](https://news.ycombinator.com/item?id=47140842)) "OpenAI Deep Research does a much better job" ([source](https://news.ycombinator.com/item?id=48288669))
- Valuable research input for LLMs (2 comments): "I almost exclusively use Deep Research as inputs to LLMs for deeper domain knowledge" ([source](https://news.ycombinator.com/item?id=45878695)) "Deep Research is invaluable" ([source](https://news.ycombinator.com/item?id=46561067))

## What people warn about

- Quality has declined over time (4 comments): "Deep Research (though I swear it was better when I could choose o3)" ([source](https://news.ycombinator.com/item?id=45904819)) "I used to like OpenAI Deep Research, but then switched" ([source](https://news.ycombinator.com/item?id=47145617)) "(OpenAI's has really gone downhill for some reason.)" ([source](https://news.ycombinator.com/item?id=47088640))
- Skips items or gives thin results (2 comments): "Repeatedly it would take shortcuts and not actually do the full list." ([source](https://news.ycombinator.com/item?id=45795269)) "got a handful of results, but nothing comprehensive" ([source](https://news.ycombinator.com/item?id=47141090))

## Alternatives hackers recommend

- [Autoresearch by Karpathy](https://hackerapproved.com/p/autoresearch.md): 37 recommend
- [Gemini Deep Research](https://hackerapproved.com/p/gemini-deep-research.md): 27 recommend
- [Kagi Research](https://hackerapproved.com/p/kagi-research.md): 9 recommend
- [LangAlpha](https://hackerapproved.com/p/langalpha.md): 8 recommend
- [Tongyi DeepResearch](https://hackerapproved.com/p/tongyi-deepresearch.md): 7 recommend

## Where to get it

- [Visit the official site](https://chatgpt.com/)

## Cite as

Hacker Approved, "OpenAI Deep Research", https://hackerapproved.com/p/openai-deep-research/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).