OpenAI has announced the rollout of GPT-5.5, featuring the strongest set of safeguards to date, which builds on the prior safety training from models like GPT-5.4. This new version is being introduced to Plus, Pro, Business, and Enterprise users across ChatGPT and Codex, with a gradual deployment strategy that emphasizes monitoring before the full API release. GPT-5.5 is noted for its agentic capabilities, enabling it to autonomously plan, utilize tools, check its work, and navigate complex tasks.

OpenAI: OpenAI is an artificial intelligence research organization dedicated to building safe and beneficial AGI through advanced language models and tools like ChatGPT and Codex. It recently launched GPT-5.5, featuring the strongest safeguards to date, which is rolling out to paid users in ChatGPT and Codex with API access planned soon. The release emphasizes agentic capabilities for complex real-world tasks and workflow automation.3334
ChatGPT: ChatGPT is OpenAI’s flagship conversational AI platform that integrates evolving GPT models for natural interactions, productivity, and specialized tasks. It is receiving the GPT-5.5 update for paid subscribers, bringing enhanced agentic features and safeguards. Recent developments focus on gradual rollouts of advanced models to improve response quality and support for real work scenarios.39
GPT-5.5: GPT-5.5 is OpenAI’s newest large language model designed as a new class of intelligence for powering agents, understanding complex goals, using tools, and completing multi-step tasks autonomously. It includes the strongest set of safeguards implemented to date to reduce misuse while enabling beneficial applications like advanced coding and research. The model is deploying to Plus, Pro, Business, and Enterprise users in ChatGPT and Codex, with Pro variant for higher tiers and API forthcoming.3335

Safeguards: GPT-5.5 incorporates OpenAI’s strongest safeguards to date, building on prior cyber-specific safety training in models like GPT-5.4.
Rollout Strategy: Gradual deployment to paid tiers in ChatGPT and Codex prioritizes monitoring before full API release.
Agentic Capabilities: The model autonomously plans, uses tools, checks its work, and persists through ambiguity on messy multi-part tasks.