Listed as deprecated in the API, so new projects should look elsewhere; existing long-running Codex setups are the main audience.

Kenning's take

What it does

GPT-5.1-Codex-Max is an OpenAI model trained for agentic coding over long tasks. It was the first OpenAI model trained to work across several context windows by compacting its history, which lets one task run over millions of tokens. It launched in Codex and later reached the API, where OpenAI now lists it as deprecated.

• Compaction to work across multiple context windows
• 400,000-token context window
• Up to 128,000 output tokens
• Extra High (xhigh) reasoning effort
• Text and image input, text output

Who it suits. Developers running long coding agent sessions in Codex or through the OpenAI API.

Key facts

Category

AI models

Part of

ChatGPT

Made by

OpenAI

Released

19 November 2025

How to get it

Launched in Codex (CLI, IDE extension, cloud and code review); in the OpenAI API through the Responses API, now listed as deprecated

Works on

API, Codex

In the UK

Yes

API price

$1.25 per million input tokens, $0.125 cached input, $10 per million output tokens

FAQ

Facts from the maker's own pages and the sources listed. Checked 2026-10-11.