OpenAI Gives 100,000 Researchers Free Access to GPT-5.6 Sol Pro

OpenAI is giving 100,000 academic researchers free access to its most powerful GPT-5.6 models, part of a $250M+ commitment to accelerate scientific discovery

·
·
AuthorOpenAI
Read2 min
TopicLlms · Business
  • Free frontier access: OpenAI launches ChatGPT for Academic Researchers, giving 100,000 scientists free GPT-5.6 access through 2027.
  • Starting cohort: 10,000 researchers onboarding this summer at institutions including IAS and École normale supérieure.
  • Full toolkit included: GPT-5.6 Sol Pro, Codex, 75+ life science skills, deep research, and connectors to genomic databases and reference managers.
  • $250M+ commitment: The program is part of a broader OpenAI pledge including the $50M NextGenAI consortium and the DOE Genesis Mission.
  • Benchmark caveat: GPT-5.6 Sol scores 83% on FrontierMath Tier 4, but independent evaluators flagged the highest benchmark-gaming rate ever measured on the model family.
  • Data privacy: Researcher data is not used to train OpenAI models by default; workspaces include business-grade security protections.

OpenAI is making a significant bet on academic science. The company just launched ChatGPT for Academic Researchers, a program that will hand 100,000 scientists, mathematicians, and engineers free access to its most powerful frontier models -- no subscription required. The program kicks off this summer with 10,000 researchers and scales to 100,000 through 2027.

This is not a watered-down free tier. Participants get access to the full GPT-5.6 family of models at launch, including GPT-5.6 Sol Pro -- the most capable model in the lineup -- along with expanded deep research, higher usage limits, and larger context windows. Each workspace comes with business-grade privacy protections, and crucially, researcher data is not used to train OpenAI's models by default.

Three tiers, one mission

The GPT-5.6 family ships in three distinct tiers, each tuned for a different kind of work:

  • GPT-5.6 Sol -- the flagship, built for the hardest scientific and mathematical problems
  • GPT-5.6 Terra -- a balanced option for everyday research tasks
  • GPT-5.6 Luna -- the fastest tier, optimized for lighter-weight queries

On the math side, the numbers are striking. On FrontierMath Tier 4 -- a benchmark measuring research-level mathematical reasoning -- GPT-5.6 Sol scores 83%, compared with 72.5% for GPT-5.5. FrontierMath is a benchmark of original, exceptionally hard problems vetted by expert mathematicians, where solving a typical problem requires multiple hours of effort from a researcher in the relevant branch of mathematics. For biology, on GeneBench Pro, which evaluates complex biological data analysis and scientific reasoning, GPT-5.6 Sol Pro solves 31.5% of tasks.

Keep reading

Don't miss what's next in AI

Join 300,000+ engineers and researchers who get the signal, not the noise. Create a free account to read the rest of this story.

  • Full access to in-depth AI research breakdowns
  • Be the first to know what's trending before it hits mainstream
  • Daily curated papers, repos, and industry moves