OpenAI rolled out GPT‑5.5 today, the newest upgrade for ChatGPT and Codex. The model promises faster understanding and more autonomous work than its March predecessor, GPT‑5.4. Developers can now tackle multi‑step tasks—planning, tool use, and self‑checking—without sacrificing speed. This move follows a rapid release cycle that began with GPT‑5.4 in March.

GPT‑5.5 shines in areas like agentic coding, computer use, knowledge work, and early scientific research. OpenAI claims the model delivers these gains while staying token‑efficient, meaning users get better results with fewer tokens compared to GPT‑5.4. Users report the model handles complex prompts with fewer iterations and reduces hallucinations during technical explanations.

API pricing jumps from $15.00 / 1M tokens output for GPT‑5.4 to $30.00 / 1M tokens for GPT‑5.5, while input and cached input rates double as well. Despite the hike, OpenAI insists the new model is more intelligent and consumes fewer tokens overall. Token efficiency translates to cost savings for high‑volume applications.

The rollout targets Plus, Pro, Business, and Enterprise users, with GPT‑5.5 Pro reserved for higher‑accuracy work. OpenAI also rolled out stricter safeguards after testing with nearly 200 trusted partners. For developers, the upgrade means faster, more reliable code generation and broader browser interaction capabilities. The update also expands Codex’s ability to navigate web apps and capture screenshots.