OpenAI just revealed what its new in-house team, OpenAI for Science, has been up to. The firm has released a free LLM-powered tool for scientists called Prism, which embeds ChatGPT in a text editor for writing scientific papers.
The idea is to put ChatGPT front and center inside software that scientists use to write up their work in much the same way that chatbots are now embedded into popular programming editors. It’s vibe coding, but for science.
Kevin Weil, head of OpenAI for Science, pushes that analogy himself. “I think 2026 will be for AI and science what 2025 was for AI in software engineering,” he said at a press briefing yesterday. “We’re starting to see that same kind of inflection.”
OpenAI claims that around 1.3 million scientists around the world submit more than 8 million queries a week to ChatGPT on advanced topics in science and math. “That tells us that AI is moving from curiosity to core workflow for scientists,” Weil said.
Prism is a response to that user behavior. It can also be seen as a bid to lock in more scientists to OpenAI’s products in a marketplace full of rival chatbots.
“I mostly use GPT-5 for writing code,” says Roland Dunbrack, a professor of biology at the Fox Chase Cancer Center in Philadelphia, who is not connected to OpenAI. “Occasionally, I ask LLMs a scientific question, basically hoping it can find information in the literature faster than I can. It used to hallucinate references but does not seem to do that very much anymore.”
Nikita Zhivotovskiy, a statistician at the University of California, Berkeley, says GPT-5 has already become an important tool in his work. “It sometimes helps polish the text of papers, catching mathematical typos or bugs, and provides generally useful feedback,” he says. “It is extremely helpful for quick summarization of research articles, making interaction with the scientific literature smoother.”
By combining a chatbot with an everyday piece of software, Prism follows a trend set by products such as OpenAI’s Atlas, which embeds ChatGPT in a web browser, as well as LLM-powered office tools from firms such as Microsoft and Google DeepMind.
Prism incorporates GPT-5.2, the company’s best model yet for mathematical and scientific problem-solving, into an editor for writing documents in LaTeX, a common coding language that scientists use for formatting scientific papers.
A ChatGPT chat box sits at the bottom of the screen, below a view of the article being written. Scientists can call on ChatGPT for anything they want. It can help them draft the text, summarize related articles, manage their citations, turn photos of whiteboard scribbles into equations or diagrams, or talk through hypotheses or mathematical proofs.
It’s clear that Prism could be a huge time saver. It’s also clear that a lot of people may be disappointed, especially after weeks of high-profile social media chatter from researchers at the firm about how good GPT-5 is at solving math problems. Science is drowning in AI slop: Won’t this just make it worse? Where is OpenAI’s fully automated AI scientist? And when will GPT-5 make a stunning new discovery?
That’s not the mission, says Weil. He would love to see GPT-5 make a discovery. But he doesn’t think that’s what will have the biggest impact on science, at least not in the near term.
“I think more powerfully—and with 100% probability—there’s going to be 10,000 advances in science that maybe wouldn’t have happened or wouldn’t have happened as quickly, and AI will have been a contributor to that,” Weil told MIT Technology Review in an exclusive interview this week. “It won’t be this shining beacon—it will just be an incremental, compounding acceleration.”
