OpenAI boosts Codex and Agents SDK for AI developers
OpenAI enhances AI development tools with updates to Codex and the Agents SDK, expanding access and adding new features.
OpenAI has announced significant updates to its Codex software engineering agent and the OpenAI Agents SDK, aiming to empower developers with enhanced tools for AI development.
Codex Updates
- Expanded Access: Previously limited to Enterprise users, Codex is now available to ChatGPT Plus subscribers, with generous usage limits (though rate limits may apply during peak demand).
- Web Connectivity: Codex can now connect to the web to install dependencies, run tests requiring external resources, and upgrade packages. Internet access remains off by default and must be enabled for specific environments. This feature rolls out first to ChatGPT Plus, Pro, and Teams users, with Enterprise access coming later.
- New Features:
- Ability to update pull requests for follow-up tasks.
- Voice dictation support for tasks.
- Support for binary files when applying patches.
- Improved error messages for setup scripts.
- Increased limits: Task diffs now support up to 5MB (from 1MB), and setup script durations extend to 10 minutes (from 5 minutes).
- Bug Fixes: Live activities for iOS Codex users have been re-enabled after resolving notification issues. Two-factor authentication is no longer required for users with single sign-on or social logins.
Agents SDK Enhancements
- TypeScript Support: The OpenAI Agents SDK now includes TypeScript, featuring guardrails, handoffs, tracing, and the Model Context Protocol for standardized third-party tool integration.
- Human-in-the-Loop Approvals: Developers can pause tool executions, serialize agent states, and approve/reject specific calls.
- Traces Dashboard: Improved visualization for voice agent runs with Realtime API session support.
- Speech-to-Speech Model: Updated model offers better instruction following, interruption behavior, and tool-calling consistency. Developers can now control speech speed.
Why It Matters
These updates reflect OpenAI’s commitment to advancing agentic AI development, making its tools more accessible and functional for developers. The improvements to Codex and the Agents SDK streamline coding workflows and enhance voice agent capabilities, positioning OpenAI as a leader in AI-driven development tools.
For more details, visit OpenAI's Codex changelog.
Related News
OpenAI o3 Outperforms GPT-5 in Complex Office Tasks Benchmark
A new benchmark reveals OpenAI's older o3 model beats the newer GPT-5 in multi-day office workflows, highlighting challenges in AI agent coordination.
GPT-5 Launches with Enhanced User Experience but Falls Short of AGI
OpenAI's GPT-5 improves ChatGPT usability and reduces hallucinations but remains far from achieving artificial general intelligence.
About the Author
Dr. Lisa Kim
AI Ethics Researcher
Leading expert in AI ethics and responsible AI development with 13 years of research experience. Former member of Microsoft AI Ethics Committee, now provides consulting for multiple international AI governance organizations. Regularly contributes AI ethics articles to top-tier journals like Nature and Science.