OpenAI is handing the keys to its most advanced technology directly to the academic world. Through a new program called ChatGPT for Academic Researchers, the company plans to provide free access to its top models for 100,000 scientists by 2027. CEO Sam Altman noted that the goal is to empower experts to solve problems in their own fields, rather than having the company try to figure out every scientific question itself.
The rollout starts this summer for ten thousand selected scientists
The initial phase begins this summer with 10,000 researchers from institutions like the Institute for Advanced Study and École Normale Supérieure. By 2027, the program will expand to the full 100,000 participants. Each person can invite up to four collaborators from their own institution to join them. Participants will get full access to frontier models like GPT-5.6 Sol Pro.
Don’t miss the best of The Mac Observer
Set us as a preferred source and our Apple reporting ranks higher in your Google Search results and Discover feed — one tap, no account changes.
This goes far beyond just typing text into a normal ChatGPT prompt window. Researchers will receive access to over 75 specific life science skills, covering areas like genetics, genomics, drug discovery, and protein modeling. The models connect directly to scientific literature, computational notebooks, genomic databases, and satellite imagery.
The company also includes Codex, its coding model, to help scientists write code, debug issues, and analyze large datasets. It even helps with the administrative side of academia, like writing grant proposals, reviewing literature, and drafting manuscripts.
High test scores meet training sessions and massive financial backing
The technology behind this initiative shows notable improvements in complex academic subjects. GPT-5.6 Sol now scores 83 percent on FrontierMath Tier 4, a benchmark test designed for research-level mathematical reasoning. This is a solid jump from the 72.5 percent achieved by the previous model. It also solves slightly more than 31 percent of tasks on GeneBench Pro, which measures complex biological data analysis.
To make sure scientists can actually use these improvements, the program provides hands-on support and training sessions. This ensures that researchers who have never used an AI tool before can still benefit from the software without being left to figure things out alone.
Applications are currently open for eligible degree-granting institutions with strong research backgrounds. This entire program falls under a larger 250 million dollar commitment the company made to support scientific research through 2027.
By removing the financial barrier to its most capable models, the company is betting that scientists will move faster when they have the right tools in hand. It puts the focus back on human curiosity, with technology acting as a quiet assistant rather than an expensive roadblock.
Discussion