Noesion Radarnoesion.ai
Release2026-07-31

DeepSeek documented how to use its models inside Codex, OpenAI's coding assistant, by pointing Codex at DeepSeek's own service through a shared configuration file. A setup script for macOS, Linux and Windows writes the required settings and backs up the existing configuration first. Only the deepseek-v4-flash model works today, with deepseek-v4-pro expected in early August 2026.

What changed

DeepSeek models were not selectable inside OpenAI's Codex coding assistant.

What it unlocks

Running DeepSeek's V4-Flash model inside the Codex command line tool, the ChatGPT desktop app and the Codex extension for VS Code, using one shared configuration file.

  • deepseek-v4-flash context window: 1,048,576 tokens
  • effective context window: 95%
  • minimum Codex client version 0.144.0

What you need to act on it

  • a DeepSeek API key
  • Codex CLI or ChatGPT desktop app installed and launched once so the ~/.codex directory exists
  • Codex client version 0.144.0 or later

Send this to someone who needs it

Shares the story and its sources. Nothing about you.

What does this mean for your job?

This is the story as everyone gets it. Once a week we send you the version written for your role — what changed, why it matters for the work you actually do, and one thing to try. Free while we tune it.