- OpenAI claims 8.4 million messages are sent weekly about advanced science and mathematics
- GPT-5.2 models can follow long chains of reasoning and verify results independently
- AI accelerates routine research tasks such as coding, literature review, and experiment planning.
OpenAI wants users to treat ChatGPT as a research collaborator, and new research claims that nearly 8.4 million messages focused on advanced scientific and mathematical topics are sent each week, generated by approximately 1.3 million users worldwide.
OpenAI highlights that this usage has grown by almost 50% over the past year, suggesting that the system is moving from occasional experimentation to regular research workflows.
These users reportedly perform work comparable to graduate studies or active research in mathematics, physics, chemistry, biology, and engineering.
Research use and integration scale.
Mathematics receives special attention in the report. GPT-5.2 models are said to sustain long chains of reasoning, verify their own work, and operate with formal testing systems like Lean.
OpenAI claims the models achieved gold-level results at the 2025 International Mathematics Olympiad and demonstrated partial success on the FrontierMath benchmark.
The report also claims that the models contributed to solutions related to Erdős’s open problems, and human mathematicians confirmed the results.
While the models do not generate entirely new mathematical theories, they recombine known ideas and identify connections between fields, accelerating formal verification and discovery of evidence.
Similar patterns appear in other scientific areas. On graduate-level benchmarks like GPQA, GPT-5.2 reportedly exceeds 92% accuracy without external tools.
Physics laboratories are reportedly using AI to integrate simulations, experimental records, documentation, and control systems, while supporting theoretical exploration.
In chemistry and biology, hybrid approaches combine general-purpose language models with specialized tools such as graph neural networks and protein structure predictors.
These combinations aim to improve reliability while keeping human oversight at the center of decision-making.
The report places these events in a broader context. Scientific progress supports medicine, energy systems, and public safety, but research often moves slowly and requires considerable manpower.
A small portion of the world’s population produces most fundamental discoveries, while projects such as drug development can take more than a decade.
OpenAI maintains that researchers are increasingly using AI tools to handle routine, time-consuming tasks, including coding, literature review, data analysis, simulation support, and experiment planning.
It cites case studies ranging from faster mathematical proofs to protein design with RetroBioSciences, where AI reportedly shortened timelines from years to months.
Although the report presents notable usage figures and comparative results, independent validation remains limited.
Questions remain about how well these results hold up over time, how widely they are applied, and whether the reported advances translate into lasting scientific advances.
These usage figures and benchmark scores stand out, but independent validation is still limited.
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
