OpenAI DevDay Unveils Major Updates for Coding, Agents, and Workflow Automation

OpenAI has unveiled a comprehensive suite of developer updates during DevDay 2026, focusing heavily on accelerating the entire software development lifecycle, from initial coding to complex refactoring and collaborative review. The core message is clear: developers can now code faster, manage parallel tasks more efficiently, and ensure their AI agents remain productive even when they step away from their physical workstation.

One of the most significant performance leaps is the introduction of Ultrafast within Codex. This enhancement drastically improves the speed of code generation, allowing developers to iterate on ideas at a pace that closely matches their natural thought process. Specifically, Ultrafast is highlighted as running up to eight times faster than the standard Astra model and four times faster than the Astra Fast variant. This speed boost is designed to eliminate the friction between an idea and its initial coded draft, enabling developers to bring complex concepts to life almost instantaneously as they type them. This level of rapid feedback loop is crucial for maintaining high productivity in modern development environments.

Complementing the speed improvements is the launch of GPT-6.1 Sol, a specialized model designed to enhance agentic capabilities. While near the performance level of the flagship Astra model, GPT-6.1 Sol is engineered for cost-efficiency and deep, complex tasks. It excels in areas like intricate refactoring, deep codebase investigations, and managing long-running agents across multiple applications. Crucially, the model offers a substantial cost advantage, providing cached input at a remarkable 95% discount compared to standard input pricing. This makes advanced, resource-intensive agentic workflows—which typically consume significant API credits—much more economically viable for enterprise adoption. GPT-6.1 Sol is positioned as the workhorse for sophisticated, multi-step coding and computer use scenarios.

Addressing the challenge of maintaining continuity in complex development projects, OpenAI introduced Codex cloud environments. This feature fundamentally changes the concept of local development dependency. By running tasks within reusable, cloud-based environments, developers are no longer tethered to their physical laptops. The environment retains all necessary components—including the entire repository structure, required dependencies, custom scripts, and specific settings—allowing agents to continue working autonomously. This means that a developer can close their laptop, trust that their agents are actively progressing on the codebase, and then resume work later, whether from a phone or another computer, ensuring zero downtime for critical development pipelines.

For developers who thrive in the command line, the Codex CLI has undergone a major refresh. This update is a substantial investment in the terminal experience, providing not only a new, improved full-screen interface and better readability but also powerful new capabilities specifically designed to manage parallel workstreams. The refreshed CLI allows users to manage multiple, concurrent tasks within a single terminal session, which is vital for complex CI/CD pipelines or multi-agent simulations. This focus underscores OpenAI's commitment to serving the power user who lives entirely within the terminal.

Furthermore, the platform has significantly enhanced collaborative and quality assurance workflows. Developers can now utilize the system to inspect pull requests directly within the ChatGPT Work or Codex desktop application. This feature allows users to view the code difference (the 'diff') and, critically, to query Codex directly about the changes. This capability moves AI assistance beyond mere generation and into the realm of active code review, providing immediate feedback on proposed modifications. Moreover, developers can leverage this system to get feedback on their own changes before they are ever shared with a team, drastically improving the quality and readiness of code submissions.

Finally, the platform's extensibility and integration capabilities were expanded through Plugin extensions. These extensions allow developers to build highly interactive panels, dedicated file viewers, and sophisticated sidebars directly within the ChatGPT interface. The support for the proposed MCP Events specification is a major architectural upgrade, enabling connected applications to trigger automated workflows and actions within ChatGPT. This deep integration means that the AI assistant is no longer just a chat interface; it is becoming a central orchestration hub for an entire suite of connected enterprise applications, solidifying its role as a comprehensive developer ecosystem tool. Overall, these updates solidify OpenAI's position as a full-stack developer partner, addressing speed, cost, persistence, and collaboration simultaneously.