OpenAI offers free ChatGPT access to 10,000 researchers, expanding to 100,000 by 2027

OpenAI launched a program called ChatGPT for Academic Researchers that will give eligible faculty and postdoctoral researchers 12 months of complimentary access to a dedicated ChatGPT workspace, starting with 10,000 users this summer and expanding to 100,000 through 2027.

The rollout, announced by OpenAI in its support documentation, aims to put the company’s latest frontier models into scientists’ hands and includes up to five complimentary seats per approved workspace and the ability for each participant to invite up to four collaborators. Applicants must be listed as an author on a paper posted to arXiv, bioRxiv, or ChemRxiv within the past three years and will verify affiliation through SheerID, according to OpenAI’s help page.https://help.openai.com/en/articles/20001406-chatgpt-for-academic-researchers

Why this matters: OpenAI is seeding university labs with its newest models — a strategy that can accelerate adoption in research communities while also serving as live validation for claims about scientific performance. The company frames the offer as a way to “accelerate discovery” across disciplines, and it positions the program alongside the GPT-5.6 family, which OpenAI says posts stronger results on scientific benchmarks than prior models.https://openai.com/index/gpt-5-6/

Who gets access and how OpenAI verifies eligibility

OpenAI’s help page details the eligibility rules: applicants must be faculty or postdoctoral researchers at a participating institution and the application typically takes one to two weeks to review. OpenAI requires a valid credit card for identity verification though the complimentary period charges a $0 checkout total; approved workspaces receive up to five complimentary seats for 12 months.https://help.openai.com/en/articles/20001406-chatgpt-for-academic-researchers

The company’s public messaging says the initial tranche is 10,000 researchers, expanding to 100,000 by 2027. Reporting from Axios and SiliconANGLE corroborates the 10,000/100,000 figure and notes that the offering is roughly equivalent in value to OpenAI’s $200/month ChatGPT Pro tier per seat, making the program a material subsidy for lab-level compute and tooling.https://www.axios.com/2026/07/29/openai-academics-research-chatgpt-sol, https://siliconangle.com/2026/07/29/openai-opens-new-chatgpt-academic-researchers-program-100000-scientists/

Why OpenAI times this with GPT-5.6 and what rivals are doing

OpenAI is rolling the program out as it promotes GPT-5.6, which its product page claims shows “broad gains across scientific research” with improved scores on benchmarks such as GPQA Diamond and FrontierMath.https://openai.com/index/gpt-5-6/

Competitors present different value propositions: Anthropic emphasises safety and long-context reasoning, while Google bundles models into Workspace and search integrations. Neither company has announced a directly comparable, large-scale complimentary researcher program in recent days, leaving OpenAI with a temporary first-mover advantage in campus research adoption.https://www.axios.com/2026/07/29/openai-academics-research-chatgpt-sol

Skepticism is warranted. The rollout is a programme announcement with limited public evidence that direct access to GPT-5.6 will produce measurable increases in reproducible scientific output. Observers also note the move serves strategic adoption goals for OpenAI — seeding labs with proprietary models can steer methodologies and downstream procurement decisions even as the company frames the effort as philanthropic.https://thenextweb.com/news/openai-chatgpt-academic-researchers-agent-harness-token-costs

OpenAI’s next concrete test will be adoption and retention metrics: how many of the initial 10,000 applicants convert to active users, and whether participating labs publish reproducible, model-assisted results. Universities and independent researchers will also be watching for details on data handling, reproducibility, and whether model-driven workflows can reliably pass peer review.