GPT-5.5 Leaves ChatGPT and Codex on October 14, but Stays in the OpenAI API
OpenAI will remove GPT-5.5 from ChatGPT, ChatGPT Work and Codex on October 14, 2026, across all plans. The company announced the change from its @ChatGPT account on September 15, as reported by Gizmochina and Business Today. Both outlets note that GPT-5.5 launched on April 23, 2026, so it will have been in those products for less than six months.
Who has to act
For most ChatGPT users nothing needs to be done: conversations move to whatever model OpenAI sets as the new default. Codex users who sign in with a ChatGPT account and have pinned GPT-5.5 in workspace defaults, saved settings, custom agents or scripts need to switch before October 14. OpenAI points them to GPT-5.6 Sol or GPT-6 Astra.
The API keeps GPT-5.5
The retirement covers plan-based products, not API keys. We checked OpenAI’s deprecations page on September 16, 2026: it lists no shutdown date for gpt-5.5. The API models that do have a removal date are older snapshots. On June 11, 2026, OpenAI scheduled gpt-5-2025-08-07, gpt-5-mini-2025-08-07, gpt-5-nano-2025-08-07, o3-2025-04-16 and several related snapshots for removal on December 11, 2026, with the GPT-5.6 family as replacements.
What GPT-5.5 costs through the API compared with its replacements
| Model | Input per 1M | Cached input | Output per 1M |
|---|---|---|---|
| GPT-5.5 | $5.00 | $0.50 | $30.00 |
| GPT-5.6 Sol | $4.00 | $0.40 | $20.00 |
| GPT-6 Astra | $10.00 | $1.00 | $50.00 |
GPT-5.5 charges $10 input and $45 output once a prompt exceeds 272K tokens. GPT-5.6 Sol is cheaper on both sides today, but its $4 / $20 rate is promotional and runs at least through November 21, 2026. OpenAI has not said what Sol will cost after that. GPT-6 Astra, released on September 3, costs twice as much as GPT-5.5 on input.
For a developer, the practical takeaway is that nothing breaks on October 14 for code that calls gpt-5.5 with an API key. Anyone moving to Sol to save money should plan for the promotion ending.
Full per-model prices, including long-context, batch and Fast mode rates, are on our OpenAI API pricing page. Every dated change we track is in the model changelog.