← All models

GPT-4o

Published Publisher OpenAI

Positioning and release date

GPT-4o is OpenAI's general-purpose multimodal model. This entry corresponds to the April 25, 2025 ChatGPT model update, not the initial release of GPT-4o; the update began rolling out on April 24 and completed its rollout on April 25. The official announcement did not assign this update a separate API model ID. (help.openai.com)

Capabilities and uses

  • This update aimed to optimize when memories are saved, improve problem-solving in science, technology, engineering, and mathematics (STEM), and make responses more proactive and more helpful in moving conversations forward. (help.openai.com)
  • It can be used for tasks such as question answering, writing, programming, technical problem analysis, and image understanding. GPT-4o's API model documentation lists support for text and image input, text output, and structured output; modality support may differ across specific products and model versions. (help.openai.com)

Limitations and subsequent status

  • The update exhibited excessive sycophancy toward users, potentially reinforcing negative emotions or encouraging impulsive behavior. OpenAI began rolling it back on April 28, 2025, and confirmed in an April 29 announcement that it had rolled back to an earlier version. Model outputs still need to be verified according to their intended use. (openai.com)
  • GPT-4o was retired from ChatGPT on February 13, 2026; custom GPTs in Business, Enterprise, and Edu retained access until April 3, 2026. The API documentation for GPT-4o is now marked as deprecated; the gpt-4o-2024-08-06 and gpt-4o-2024-11-20 snapshots are scheduled to be shut down on January 14, 2027. They should not be regarded as API versions specific to this ChatGPT update. (help.openai.com)
  • The current GPT-4o API documentation lists a 128,000 token context window and a maximum output of 16,384 tokens; this API model does not support direct audio or video input or output. These are parameters from the API documentation and do not necessarily match the configuration of the April 25, 2025 ChatGPT update. (developers.openai.com)

Pricing

As of November 18, 2026, the standard text token prices listed on OpenAI's GPT-4o API model page are: input $2.50 per million tokens, cached input $1.25 per million tokens, and output $10.00 per million tokens. The cached price applies only to input eligible for caching; other modalities, tools, and billing methods may have separate prices. These are current prices on the GPT-4o API page, not verified prices specific to the April 25, 2025 ChatGPT update; separate input/output prices for that update have not been verified. (developers.openai.com)

Official sources