← ChatGPT & Codex changelog

OpenAI API — May, 2026

This major release delivers significant advancements in AI models, including new Realtime voice models for speech-to-speech agents, streaming translation, and transcription, alongside a chat-latest snapshot providing access to the newest Instant model. Enterprise users benefit from enhanced security through workload identity federation, Secure MCP Tunnel, new Admin API capabilities for comprehensive management, and multiple IP allowlists. Additionally, prompt caching is extended by default, and important model deprecations for DALL·E 2/3 and the Realtime API Beta require migration to newer alternatives like gpt-image models and the released Realtime API.

Added

  • Released chat-latest snapshot which points to the latest Instant model currently used in ChatGPT. We recommend leveraging GPT-5.5 for production API usage, but feel free to use this model to test the latest improvements for chat use cases. The underlying model snapshot will be regularly updated. Read more here.
  • Released workload identity federation. Trusted workloads can exchange externally issued identity tokens for short-lived OpenAI access tokens without storing long-lived API keys.
  • Added new Admin API capabilities for managing spend alerts, model allowlists, data retention settings, and hosted tool permissions, plus querying granular billing line items.
  • Released Secure MCP Tunnel for enterprise customers. Secure MCP Tunnel lets supported OpenAI products including ChatGPT web, Codex, Responses API, and AgentKit connect to private or on-prem MCP servers through a customer-hosted tunnel-client without exposing those servers to the public internet.
  • Added return_token_budget for the Responses API web search tool. Use it to opt in to longer GPT-5+ reasoning web search runs for high-effort research and evaluation workloads.
  • Released GPT-Realtime-2, a new realtime voice model with configurable reasoning for speech-to-speech agents, along with GPT-Realtime-Translate for streaming speech translation and GPT-Realtime-Whisper for streaming speech-to-text.
  • Released the OpenAI Developers plugin for Codex. This helps you build AI applications and agents in Codex with OpenAI Platform access and OpenAI API setup guidance.
  • Released chat-latest snapshot which points to the latest Instant model currently used in ChatGPT. We recommend leveraging GPT-5.5 for production API usage, but feel free to use this model to test our latest improvements for chat use cases. The underlying model snapshot will be regularly updated. Read more here.

Changed

  • For organizations without ZDR enabled, prompt_cache_retention now defaults to 24h instead of in_memory, enabling extended prompt caching by default. Learn more.
  • You can now manage multiple IP allowlists and apply each one at the project level or across the whole organization. To configure them, go to Settings > Security > IP allowlist.
  • DALL·E model snapshots dall-e-2 and dall-e-3 were deprecated and removed from the API on May 12, 2026. We recommend using gpt-image-2, gpt-image-1, or gpt-image-1-mini instead.
  • The Realtime API Beta was deprecated and removed from the API on May 12, 2026. If you are still using the beta interface, migrate to the released Realtime API. See the migration guide and the full deprecations page.
  • Updated the Realtime and audio guide, added a dedicated Realtime translation guide, refreshed Realtime transcription for streaming transcripts, and moved realtime prompting guidance into Using realtime models.
  • The updated Agents SDK is now available in TypeScript, with support for sandbox agents and an open-source harness built in. Learn more here.
  • Admin APIs are now supported in the OpenAI SDKs for Node, Python, Go, Ruby, and Java. See the Admin APIs guide for setup instructions and examples.

Removed

  • Deprecated DALL·E model snapshots and the Realtime API Beta.