OpenAI expands ChatGPT with new connectors; Altman clarifies GPT-5 usage
OpenAI is diligently working to stabilize the rollout of its flagship large language model, GPT-5, following what CEO Sam Altman candidly described as a “bumpy” debut. The company is now prioritizing its underlying infrastructure and refining usage limits to better serve its vast user base, which includes 700 million active weekly ChatGPT users. This strategic pivot follows an initial period where some users expressed a preference for older models, prompting OpenAI to reinstate access to GPT-4o.
Altman recently outlined OpenAI’s immediate strategy for allocating computing resources amidst increased demand for GPT-5. The company’s primary focus is to ensure that current paying ChatGPT users receive greater overall usage capacity than they did prior to GPT-5’s release. While specific figures for this increase were not provided, Altman had previously indicated a trial usage limit of 3,000 messages per week for ChatGPT Plus subscribers, who pay $20 per month, when utilizing GPT-5’s more advanced “thinking” mode—a feature designed for deeper reasoning on complex problems. Intriguingly, reports from AI app creators suggest a notably lower manual usage cap for this “thinking” mode on the ChatGPT Team plan, priced at $30 per user per month, set at just 200 messages per week.
Beyond direct user access, OpenAI is also adjusting the availability of GPT-5 through its Application Programming Interface (API) for third-party developers. Altman stated that the company would prioritize API demand up to currently allocated capacity and existing customer commitments. This means that developers with established contracts or existing API access will receive priority for GPT-5 integration, while others may face longer wait times. OpenAI anticipates supporting approximately 30% additional new API growth from its current standing. With five million businesses reportedly paying for access to its ChatGPT services, the company already serves a substantial developer ecosystem. To further alleviate capacity constraints and enhance performance for both ChatGPT and API users, OpenAI plans to roughly double its compute fleet over the next five months. The precise scale or technical specifications of this infrastructure expansion were not detailed.
In addition to resource management, OpenAI has expanded ChatGPT’s integration capabilities. Subscribers to ChatGPT Plus can now connect the application to search for files and projects across various third-party platforms, including Box, Canva, Dropbox, HubSpot, Notion, Microsoft SharePoint, and Microsoft Teams. This allows users to seamlessly search their connected accounts directly within a ChatGPT conversation, eliminating the need to toggle between separate applications. For the higher-tier ChatGPT Pro subscribers, priced at $200 per month, new connectors for Microsoft Teams and GitHub have been introduced. These additions complement OpenAI’s existing integrations with services such as Gmail, Google Drive, and Google Calendar. Users must manually enable these external account connections through ChatGPT’s settings. A notable regional limitation, however, is that these new connectors are currently unavailable for Plus and Pro subscribers in Europe, Switzerland, and the United Kingdom. Furthermore, while the connectors are in beta, they are disabled by default for Enterprise and Education plans, though administrators have the option to enable them.
By strategically combining capacity planning with new productivity integrations, OpenAI is positioning GPT-5 not merely as a more powerful AI model, but as an integral component of a more connected digital workspace. The phased approach to compute allocation underscores the company’s commitment to serving its existing customer base while simultaneously scaling to meet future demand. As the planned compute expansion comes online, paying users are expected to be the primary beneficiaries, gaining both increased availability and more versatile ways to embed ChatGPT into their daily workflows. However, the immediate priority for OpenAI remains ensuring the stable and smooth operation of GPT-5 for all users who wish to leverage its capabilities.