AI firm OpenAI (PC:OPAIQ), which is backed by tech giant Microsoft (MSFT), has introduced a new benchmark called FrontierScience to measure how well AI models can reason and understand advanced science topics like biology, chemistry, and physics. According to a blog post, the questions in the benchmark were created and reviewed by experts and are meant to be challenging, original, and meaningful.
Claim 50% Off TipRanks Premium and Invest with Confidence
- Unlock hedge-fund level data and powerful investing tools designed to help you make smarter, sharper decisions
- Stay ahead of the market with the latest news and analysis so your portfolio is always positioned for maximum potential
Importantly, the FrontierScience benchmark is split into two parts: one that tests Olympiad-style reasoning, like the type found in academic competitions, and another that evaluates real-world research skills. OpenAI’s newest model, GPT-5.2, performed well by scoring 77% on the Olympiad section and 25% on Research. Other AI models followed behind:
- Claude Opus 4.5 from Anthropic (PC:ANTPQ) scored 71.4% and 17.5%
- Google’s (GOOGL) Gemini Pro 3 earned 76.1% and 12.4%
- Grok 4 from xAI (PC:XAIIQ) scored 66.2% and 15.9%.
OpenAI says that improving scientific reasoning in AI will require better overall thinking skills and more targeted efforts in science-related tasks. Nevertheless, the company believes benchmarks like FrontierScience are useful for identifying what current AI systems struggle with, so developers can focus on fixing those weak spots. In the long run, OpenAI hopes these improvements will help AI become a trustworthy partner for scientific research and discovery.
Is MSFT Stock a Good Buy?
Turning to Wall Street, analysts have a Strong Buy consensus rating on MSFT stock based on 32 Buys and two Holds assigned in the last three months. Furthermore, the average MSFT price target of $632.22 per share implies 33.2% upside potential.
