What Codex Is and Why 5 Million Users Matter
Codex is OpenAI’s coding agent that combines large language models with software tools to generate, debug and refactor code, understand entire repositories, and automate repetitive engineering workflows for individual developers and enterprise teams. The service has moved from research preview to a central pillar of OpenAI’s enterprise coding AI strategy, and its adoption curve shows how quickly coding assistants are becoming standard in development environments. Product head Tibo Sottiaux announced that Codex has crossed 5 million users and that OpenAI is resetting usage limits to celebrate, a move that encourages heavier experimentation and helps cement habits. Codex’s rise is striking because Anthropic’s Claude Code reached market earlier and built strong momentum, yet Codex has still scaled from launch in May 2025 to millions of active users in a little over a year, driven by both individual developers and large organizations.
Windows Support Brings Computer Use to Mainstream Enterprise Desktops
The latest Codex update (version 26.527) brings Computer Use to Windows, a critical step for enterprise adoption where Windows remains the default desktop. Codex can now see, click and type across Windows desktop apps in the foreground, automating tasks such as running build tools, editing project files or managing internal systems without custom integrations. On Windows, Computer Use controls the active desktop session rather than working in the background, so users will see the pointer move and windows change as tasks run. For longer jobs, teams can keep devices unlocked and connected, then monitor and steer Codex from the ChatGPT mobile app or from Codex on Mac, turning the PC into a persistent host for project files and local context. Computer Use is not yet available in some regions, but the Windows rollout sharply widens the practical footprint of Codex inside corporate development stacks.
Amazon Bedrock Integration Pushes Codex Deeper into Enterprise Infrastructure
OpenAI has made Codex, alongside GPT-5.5 and GPT-5.4, available on Amazon Bedrock, indicating a shift from single-cloud dependence to broader infrastructure partnerships for enterprise coding AI. Customers can now access Codex through their existing AWS commitments, reducing procurement friction and aligning AI spending with established cloud governance, security and compliance controls. According to TheElec, the move was enabled by a revision of OpenAI’s partnership agreement with Microsoft, which had previously limited distribution of OpenAI models to Azure. For enterprise teams, Codex on Amazon Bedrock means they can standardize on AWS tooling while tapping Codex for agent-based coding, repository-scale reasoning and integration with other Bedrock services. This also strengthens OpenAI’s competitive stance against Claude Code, which is widely used in terminal-centric workflows, by embedding Codex inside the same cloud environments where large engineering organizations already build and run their production systems.
Education and Workforce Pipelines: Codex Beyond Developers
OpenAI’s partnership with Carahsoft highlights how Codex is expanding beyond pure developer audiences into operational and education teams, which are key to long-term workforce development. A joint webinar titled “Building With Codex in Education” is aimed at colleges, universities and other institutions, with explicit emphasis on internal workflows rather than classroom-only use. Nicole Carter, GTM for Education at OpenAI, says the session will show how education teams can turn everyday institutional needs into practical tools and workflows, from reducing manual effort to quickly prototyping internal solutions. The event targets both technical and non-technical staff, reinforcing Codex as a general-purpose workflow engine, not only a coding assistant. Carahsoft’s role as a distributor to public sector organizations means Codex is being positioned as a standard component in administrative and operational stacks, seeding familiarity among future workers and non-engineering teams who will increasingly collaborate with coding agents.
Rate Limit Resets and the Enterprise Coding AI Battle with Claude
OpenAI’s decision to reset Codex rate limits after hitting 5 million users suggests a strategy focused on retention and deeper engagement rather than short-term constraint. By removing accumulated ceilings, OpenAI invites users to push Codex into larger projects and team-wide workflows, which can translate into higher-value enterprise deployments. Codex has already shown strong traction among companies such as Cisco, Nvidia, Ramp, Rakuten and Harvey, where reports include sharply reduced code review times and compressed project cycles. At the same time, Anthropic’s Claude Code still leads in revenue and mindshare as a terminal-native coding agent. The contest is shifting from whose model writes better functions to whose platform fits more cleanly into enterprise environments: Windows desktops, AWS cloud stacks, public sector operations and education workflows. With Codex Windows support, Codex Amazon Bedrock availability and education-sector initiatives, OpenAI is building a broad, infrastructure-first answer to Claude’s early lead.