OpenAI releases GPT-5.2 models for scientific research applications
Get Alerts MSFT Hot Sheet
Join SI Premium – FREE
OpenAI announced the release of GPT-5.2 Pro and GPT-5.2 Thinking, positioning them as models designed for scientific and mathematical applications. The company stated these represent improvements over previous versions in mathematical reasoning and scientific workflows.
According to OpenAI's benchmarks, GPT-5.2 Pro achieved 93.2% accuracy on GPQA Diamond, a graduate-level science question benchmark covering physics, chemistry, and biology. GPT-5.2 Thinking scored 92.4% on the same test. On FrontierMath, which evaluates expert-level mathematics problems, GPT-5.2 Thinking solved 40.3% of Tier 1-3 problems, compared to 31.0% for the previous GPT-5.1 Thinking model.
The company highlighted a case study where GPT-5.2 Pro contributed to resolving an open research problem in statistical learning theory. The model helped produce a proof addressing whether collecting more data reliably improves results in maximum likelihood estimation scenarios. OpenAI stated that researchers asked the model to solve the problem directly, then verified the proof through external expert review.
OpenAI said the models demonstrated capabilities in mathematical reasoning, multi-step logic, and maintaining consistency across complex analyses. The company indicated these tools could assist researchers in coding, data analysis, and experimental design while emphasizing that human verification and domain expertise remain necessary.
The announcement follows OpenAI's publication of a paper last month documenting case studies of GPT-5 applications across mathematics, physics, biology, computer science, astronomy, and materials science. Microsoft (NASDAQ: MSFT) maintains a partnership with OpenAI and has invested billions in the artificial intelligence company.
Serious News for Serious Traders! Try StreetInsider.com Premium Free!
You May Also Be Interested In
- Citigroup set to join top banks in Anthropic IPO - Bloomberg
- Saba Capital funds reapprove merger plan targeting Q4 2026 close
- Abbott settles preterm formula NEC lawsuits for $670 million
Create E-mail Alert Related Categories
Corporate NewsRelated Entities
Maynard Um, Mark Zuckerberg, ARKSign up for StreetInsider Free!
Receive full access to all new and archived articles, unlimited portfolio tracking, e-mail alerts, custom newswires and RSS feeds - and more!



Tweet
Share