OpenAI has launched a new version of its ChatGPT deep research tool, designed to be more cost-effective. This “lightweight” model is now accessible to both free users and subscribers across various paid plans.
The lightweight deep research tool is based on a modified version of OpenAI’s o4-mini model. According to the company, it is “almost as intelligent” as the original deep research model, which is known for generating extensive reports complete with citations and process summaries. OpenAI claims that the lightweight version offers “shorter” responses without compromising on the depth and quality users expect.
While the traditional deep research model remains exclusive to paying customers, the lightweight version is currently being made available to free users, who can utilize it for up to five tasks per month. Paid subscribers will also benefit from this model, which will automatically activate when they reach their original research limits. Team and Plus users will be able to perform a total of 25 tasks monthly across both models, while Pro users can access up to 250 tasks. Enterprise and Education users are set to gain access to the new version next week, with usage limits similar to those of Team and Plus users.
It’s noteworthy that OpenAI emphasizes the new version is “significantly cheaper” to operate, which is crucial in a context where the company incurs “tens of millions of dollars” just to process polite user interactions such as saying please and thank you.
