OpenAI to merge ChatGPT and Codex under Greg Brockman's oversight

OpenAI is consolidating ChatGPT, Codex, and developer API under one team. Greg Brockman, the company's co-founder and president, is taking full control of product strategy in an attempt to avoid fragmentation that has slowed development.

A unified agentic platform instead of a patchwork

In an interview with Wired, Brockman explained the logic behind the merger: OpenAI is investing in a unified agentic platform where ChatGPT (consumer-facing) and Codex (developer-facing) will become interfaces to a single foundation. Until now, these products have evolved in parallel, each with its own product team, its own roadmap, and often divergent visions for the future. Now everything is under one roof, with a single team bearing full responsibility from idea to production-ready solution. This enables faster synchronization of development:

• Synchronize base model updates between consumer and developer API simultaneously
• Avoid situations where one product begins to lag behind the other in functionality
• Share insights and learnings between development directions more quickly
• Simplify architecture and reduce technical debt that accumulates from duplicated logic

Brockman uses an interesting phrase: "side quests are over." Previously at OpenAI there were too many parallel projects and experiments (Whisper for speech, DALL-E for images, GPTs for customization, Canvas for integration) that distracted from the main goal and blurred focus.

From experiments to a unified development vector

Over the past three years, OpenAI has released many diverse things: from speech input (Whisper) and image generation (DALL-E) to custom GPTs and Canvas integration. Each project required its own resources, teams, and constant attention. But the company's main development vector remained clear—agents. Agents are systems that make decisions and take action without human intervention: they click on websites, fill out forms, write and run code, search for information. They differ from regular chat interfaces in that they can perform multi-step tasks autonomously, double-checking results and adapting to unexpected obstacles.

The consolidation of ChatGPT and Codex under one team signals that an agent-oriented approach is not just a new feature, but an architectural decision at the company level. An agent-first approach requires deep alignment: an agent working in Consumer ChatGPT must have access to the same set of tools and logic as an agent in Developer API. Codex was the ideal entry point for developers who want to embed automatic code generation in their applications and services. But it developed separately, had a separate owner, a separate roadmap. This slowed innovation and created confusion in the ecosystem.

What this means for the ecosystem

For developers, the consolidation could significantly simplify the integration process: instead of different APIs and interfaces for ChatGPT and Codex, there will be a unified platform, one documentation, one set of abstractions for working with agents. For competitors (Anthropic, Google, Meta) this is another clear signal that OpenAI has chosen a strategic vector instead of scattering resources. Anthropic has Claude and Claude API, but there is no clear unification around agents. Google has Gemini API and MakerSuite, but the structure remains fragmented. Meta invests in LLaMA, but there is no unified agent frontend yet. OpenAI has chosen the path of focus: instead of many parallel experiments—one clear priority. And that priority is agents that work everywhere.