OpenAI has implemented significant price reductions on its GPT-5.6 models, slashing Luna by 80% and Terra by 20% as cost-conscious businesses grow more cautious about ballooning AI bills. According to recent reports, Luna's input price fell to 20 cents per million tokens from $1, while its output price dropped to $1.20 from $6. Terra's rates fell to $2 and $12 per million tokens, down from $2.50 and $15 respectively. The cuts, implemented on July 30, reflect mounting pressure from cost-conscious enterprises and competition from cheaper Chinese rivals, including Moonshot AI's Kimi K3 and Z.ai's GLM-5.2. OpenAI framed the move as an efficiency gain, stating that their strategy remains focused on advancing both capability and efficiency so each generation of intelligence can accomplish more work at a lower cost.
OpenAI has launched 'ChatGPT for Academic Researchers', a new programme that will provide free access to the company's frontier AI models to 100,000 scientists, mathematicians and engineers by 2027. According to an OpenAI blog post, the programme will begin this year with 10,000 researchers before gradually expanding to the full capacity over the next two years. The initiative is initially rolling out this summer at prestigious academic institutions including the Institute for Advanced Study and École normale supérieure. The programme forms part of OpenAI's broader commitment of more than $250 million through 2027 to support external scientific research and discovery, aiming to democratize access to cutting-edge AI tools and accelerate scientific progress. As OpenAI notes, researchers are already using AI models to sift through data and write grants — this just makes the relationship a bit more formal.
Researchers will receive free access to OpenAI's frontier models across ChatGPT, ChatGPT Work and Codex, including the GPT-5.6 family of models at launch. As reported by OpenAI, GPT-5.6 Sol scores 83 per cent on FrontierMath Tier 4, a benchmark that evaluates research-level mathematical reasoning, compared with 72.5 per cent for GPT-5.5. On GeneBench Pro, which measures scientific reasoning and complex biological data analysis, GPT-5.6 Sol Pro solves 31.5 per cent of tasks. Participants will also have access to more than 75 life sciences skills spanning genetics, genomics, sequencing, single-cell analysis, protein modelling and drug discovery. Researchers can use these models for tasks ranging from genomic analysis and protein modeling to literature reviews and grant writing. While the least charitable read of the program is that OpenAI is looking for new sources of training data, the company says that by default, researchers' data will not be used to train models to ensure research integrity.
The programme includes expanded Deep Research capabilities, higher usage limits and larger context windows to support scientific reasoning and long-form research workflows. According to OpenAI, researchers will receive connectors providing access to scientific literature, public genomic and clinical databases, satellite imagery, computational notebooks, data platforms and reference managers. Codex will assist with writing and debugging code, analysing datasets and building reproducible workflows. ChatGPT Work will support longer-term projects such as conducting literature reviews, identifying funding opportunities, preparing grant applications, drafting manuscripts and communicating research findings. Each workspace includes business-grade privacy and security protections, and data is not used to train models by default to ensure research integrity. This isn't the first time the company has courted researchers — OpenAI introduced Prism in January, an AI-powered tool for working with scientific journals and documents, available to anyone with a ChatGPT account.
The programme will offer training tailored to different levels of expertise, ranging from introductory sessions on improving research workflows to advanced applications of AI in scientific research. As reported by OpenAI, participants will receive support from specialists familiar with research workflows, including assistance with integrating these tools into their work. Depending on the programme's progress and adoption, OpenAI plans to create more opportunities for researchers across disciplines to share practical approaches and learn from one another. The initiative aims to accelerate scientific progress by providing researchers with the tools they need to tackle complex problems and improve research productivity across multiple domains. OpenAI's version of ChatGPT for schools, ChatGPT Edu, follows a similar logic, with the company noting that making more scientific fields dependent on their tools could also pave the way for future revenue from for-profit research.
The initial programme is open to qualified researchers at selected academic institutions that are recognized, degree-granting colleges or universities with a high level of research activity. According to OpenAI, approved researchers may invite up to four collaborators from their institution, with each collaborator requiring separate verification of institutional affiliation. While eligible researchers will receive free access through the new programme, ChatGPT is available through multiple paid subscription plans ranging from ₹0 for the Free plan to ₹19,900 per month for the Pro plan for general users. The programme represents OpenAI's commitment to making its most advanced AI models available at no cost to 100,000 academic researchers globally.
OpenAI reported that approximately 1.3 million people use ChatGPT for advanced science and mathematics each week, generating about 8.4 million messages. In mathematics, papers acknowledging ChatGPT's contribution on arXiv increased from 14 in February to 100 in the first three weeks of July. Researchers in the top 20% of AI usage within their field submit requests for tasks estimated to require four hours or more at nearly twice the rate of their peers, at 7% compared with 3.5%. The initiative also includes NextGenAI, a $50 million initiative supporting research institutions, and work with the Department of Energy's Genesis Mission to bring frontier AI to researchers at national laboratories and universities. The pricing cuts could potentially ease cost concerns for researchers while strengthening OpenAI's growth story ahead of anticipated IPOs, though analysts note that shrinking per-token revenue on already thin-margin inference businesses may complicate future profitability demonstrations.
