OpenAI Adds 50,000 More ChatGPT Grants for Researchers, but Swaps Pro for Plus
OpenAI is adding 50,000 free ChatGPT grants for academic researchers, but instead of the top-tier Pro plan, it is offering the cheaper Plus plan. One grant covers the researcher and up to two colleagues, so up to 150,000 people in total. Applications open on November 23 and will be processed on a first-come, first-served basis, with no lottery. The company explains the switch by saying most pilot participants would have been fine with Plus limits, and the savings let it reach more teams.
AI-processed from TNW; edited by Hamidun News
OpenAI plans to scale up its ChatGPT for Academic Researchers program: the company will give away 50,000 more free grants to scientists. But the generosity comes with a catch. New participants will get the cheaper Plus plan, not the top-tier Pro plan that pilot participants had. For university labs deciding which tools to budget for, this is an important signal: free access to AI is becoming mass-market, but it is scaled back.
How the grants will be handed out
Applications open on November 23. Previously, spots were effectively raffled off; now, grants will go to the first eligible applicants. If you work at a university or research institute, it is worth applying in the first few days rather than waiting for colleagues to snap them all up.
Each grant covers the researcher and up to two of their colleagues. If all 50,000 grants are used at maximum capacity, up to 150,000 people will gain access. The pilot's scheme was more generous: each scientist could invite up to four co-authors. In other words, OpenAI is widening reach while shrinking the size of each "team" by one grant.
The program's new goal is 200,000 researchers, twice what the company set when it launched the pilot. The program currently has 20,000 scientists from 33 countries.
Why Plus instead of Pro
Here is the most notable caveat. Pilot participants used Pro, ChatGPT's highest tier. The new grants will include Plus. OpenAI explains that most pilot participants would have been fine with Plus limits, and the switch lets it cover more teams.
The logic is clear. Pro is for people running heavy, long tasks who hit their limits. An ordinary scientist who reads papers, edits text, or asks for an explanation of someone else's method usually gets by with a basic paid plan. In practice, though, researchers with heavy workloads, such as processing large volumes of literature, may hit the ceiling and have to pay for an upgrade themselves. Check Plus limits against your real tasks before building a lab workflow around it.
What the usage study showed
Alongside the announcement, OpenAI published an analysis of how more than 1,800 scientists from the pilot used ChatGPT. On average, reading and learning made up 28% of their research messages, and writing made up 27%. In other words, AI most often works as an assistant that helps people get up to speed on others' work and draft their own, rather than as an autonomous "discoverer."
Survey respondents said that without AI, the same work would have taken nearly twice as long. But this figure has an important limitation: it is based on self-assessment, not measurement. OpenAI itself acknowledges that the study does not prove AI speeds up scientific discovery. The company also names the real constraints on testing new ideas: funding, a shortage of staff, and lab capacity. A neural network will not replace equipment or graduate students.
Why OpenAI is doing this
The grants are one of several science initiatives from the company. OpenAI already offers free seats on its biology model to American university labs, and it has committed $5 million to the Simons Foundation for a project to help mathematicians verify AI results.
Competitors are also courting scientists. Anthropic launched an AI environment for labs in June, and Google DeepMind is opening its scientific models to researchers. Universities are future users, paper authors, and a source of reputation: if an AI tool becomes routine during graduate school, it stays with a person long after.
The practical takeaway is simple. If you are a research scientist, apply on November 23 and plan around Plus, not Pro. If you build tools for scientists, pay attention to query structure: reading, learning, and writing account for more than half of research messages, and those are the scenarios to make easiest to use.
- Nov 21, 2026Developer Skips Pinecone: Searching 72,450 Code Chunks in 50 ms
- Nov 21, 2026Gemini 4 Neon shrank the TS Rust binary from 47.44 to 7.97 MB in under 10 hours
- Nov 21, 2026FDE job postings grew 4.6x in a year: what a forward deployed engineer actually does at the client's site
- Nov 21, 2026A swimming team captain handed ChatGPT three scheduled tasks and cut her routine work by a multiple
Key Questions Answered
How the grants will be handed out?
Applications open on November 23. Previously, spots were effectively raffled off; now, grants will go to the first eligible applicants. If you work at a university or research institute, it is worth applying in the first few days rather than waiting for colleagues to snap them all up. Each grant covers the researcher and up to two of their colleagues. If all 50,000 grants are used at maximum capacity, up to 150,000 people will gain access.
Why Plus instead of Pro?
Here is the most notable caveat. Pilot participants used Pro, ChatGPT's highest tier. The new grants will include Plus. OpenAI explains that most pilot participants would have been fine with Plus limits, and the switch lets it cover more teams. The logic is clear. Pro is for people running heavy, long tasks who hit their limits. An ordinary scientist who reads papers, edits text, or asks for an explanation of someone else's method usually gets by with a basic paid plan.
What the usage study showed?
Alongside the announcement, OpenAI published an analysis of how more than 1,800 scientists from the pilot used ChatGPT. On average, reading and learning made up 28% of their research messages, and writing made up 27%. In other words, AI most often works as an assistant that helps people get up to speed on others' work and draft their own, rather than as an autonomous "discoverer." Survey respondents said that without AI, the same work would have taken nearly twice as long.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.