Codex CLI adds a built-in Google Cloud Vertex AI provider
Version 0.165.0 lets you point OpenAI's coding agent at Vertex AI with one config line, alongside GPT-6.1 Terra support and memory v2 by default.
OpenAI shipped Codex CLI 0.165.0 on 27 October, adding a built-in Google Cloud Vertex AI provider. Setting model_provider = "gcp" in config is now enough to route the coding agent through Vertex instead of the OpenAI API directly.
The same release adds GPT-6.1 Terra to the model catalogue, turns on memory v2 by default for fresh profiles (with support for importing memories from other agents), and enables mid-conversation reasoning-effort changes by default for models that support it. The session picker also gained a repository filter and directory grouping, useful once you’re juggling several linked worktrees.
model_provider = "gcp"
Why it matters: teams already committed to Google Cloud for billing, procurement or data residency can now run Codex against Vertex-hosted models without a separate OpenAI account or a custom proxy in front of the CLI.
The caveat: this only helps if your organisation already has Vertex AI access set up. For everyone else, it’s one more provider option rather than a reason to switch.