OpenAI GPT-4o

Superseded (verified 2026-10-02)

GPT-4o (launched 2024-05-13) was retired from ChatGPT on 2026-02-13 together with GPT-4.1, GPT-4.1 mini and o4-mini (OpenAI help article “Retiring GPT-4o and other ChatGPT models”; the page returned 403 on direct fetch, date confirmed in the article’s search listing and several secondary sources). In the API, the chatgpt-4o-latest snapshot was shut down 2026-02-17; the gpt-4o text/vision model itself is still listed on OpenAI’s models page with no deprecation entry. The gpt-4o-audio* and gpt-4o-realtime* legacy models shut down on 2027-01-20 (announced 2026-07-20; the earlier preview variants ended 2026-05-07). Successors: gpt-5-and-gpt-6-family (current flagship line); for the intermediate generation see gpt-4-dot-1.

OpenAI GPT-4o

by OpenAI

“Omni” multimodal flagship of May 2024: one network handling text, image and audio input, text output via the API (audio output through the dedicated audio/realtime models), about 2x faster and 50% cheaper than GPT-4 Turbo at launch.

Specification (API model page, accessed 2026-10-02)

• Context window 128,000 tokens; max output 16,384; knowledge cutoff 2023-10-01; default snapshot gpt-4o-2024-08-06.
• Usage-based API pricing per token, with discounted cached input (unchanged from the 2024 note). Current prices: see https://developers.openai.com/api/docs/models/gpt-4o
• Input: text and images; output: text. Supports streaming, structured outputs, function calling, file search, web search, Batch API and fine-tuning; endpoints Chat Completions, Responses, Assistants, Batch.
• Launch claims (2024): audio response in as little as 232 ms, 320 ms on average (OpenAI, May 2024).

What it was good at

Vision tasks (document and screenshot reading, charts, scene description), real-time voice interaction, and cheap high-volume multimodal workloads such as invoice extraction, image moderation and visual Q&A. The same use cases are now served by newer models in gpt-5-and-gpt-6-family and by competitors.

Lifecycle summary

Date	Event
2024-05-13	Launch in ChatGPT and API
2026-02-13	Retired from ChatGPT (with GPT-4.1, GPT-4.1 mini, o4-mini)
2026-02-17	chatgpt-4o-latest API snapshot shut down (replacement gpt-5.1-chat-latest)
2026-05-07	gpt-4o-realtime-preview and gpt-4o-audio-preview models shut down
2027-01-20	gpt-4o-audio, gpt-4o-realtime and mini variants scheduled to shut down

Related

• OpenAI
• gpt-4-dot-1
• gpt-5-and-gpt-6-family
• generative-pretrained-transformer

Sources

• https://developers.openai.com/api/docs/deprecations (accessed 2026-10-02)
• https://developers.openai.com/api/docs/models/gpt-4o (accessed 2026-10-02)
• https://help.openai.com/en/articles/20001051-retiring-gpt-4o-and-other-chatgpt-models (403 on fetch; search listing and secondary reports e.g. https://www.nxcode.io/resources/news/gpt-4o-retirement-2026, accessed 2026-10-02)
• https://en.wikipedia.org/wiki/GPT-4o (launch date and latency claims; accessed 2026-10-02)