Why does `openclaw models list` show `configured,missing` for `gpt-5.4`?

This means OpenClaw recognizes the model exists in your configuration, but the runtime system cannot find or connect to it. This is often due to the model not being fully implemented or deployed in the specific OpenClaw version you are using.

Can I force `gpt-5.4` to work on OpenClaw 2026.3.2?

No, not without patching OpenClaw itself. The issue is a release gap. The recommended solutions are to upgrade to a newer OpenClaw version that includes the fix or to use a previously supported model like `gpt-5.3-codex`.

How do I check if `gpt-5.4` is supported in my OpenClaw version?

Run `openclaw models status --probe`. If it shows `missing` or `Unknown model`, it's not supported in your current runtime. You can also consult the OpenClaw release notes or the GitHub repository for support status.

Fixing OpenClaw's openai-codex/gpt-5.4 Unknown Model / HTTP 404 Error

TL;DR — Quick Fix

The `openai-codex/gpt-5.4` model is not fully supported in OpenClaw 2026.3.2 due to a release gap; upgrade to a newer version or wait for a patch.

Run Diagnostics

Next Step

Fix now, then reduce repeat incidents

If this issue keeps coming back, validate your setup in Doctor first, then harden your config.

Open Doctor Harden Config

Error Signal

FailoverError: Unknown model: openai-codex/gpt-5.4FailoverError: HTTP 404: 404 page not foundopenai-codex/gpt-5.4 ... configured,missing

What's Happening

You've configured openai-codex/gpt-5.4 or openai-codex/gpt-5.4-codex in your OpenClaw 2026.3.2 setup. It shows up when you run openclaw models list, appearing as configured,missing. However, when you try to use it, you get runtime errors like FailoverError: Unknown model: openai-codex/gpt-5.4 and FailoverError: HTTP 404: 404 page not found.

The Fix

This issue stems from a release gap. The fix is not in your local configuration, but in the OpenClaw version.

Option 1: Upgrade OpenClaw

The support for gpt-5.4 via openai-codex was merged to the main branch via commit 5d4b04040. Ensure you are running a version of OpenClaw that includes this commit or later. The problem appears to be resolved in newer development builds.

Option 2: Use a Supported Model

As a temporary workaround on 2026.3.2, revert to a known working model like openai-codex/gpt-5.3-codex for your sessions.

Why This Occurs

While the model configuration was accepted, the runtime logic and API integrations for openai-codex/gpt-5.4 were not fully complete or deployed in the 2026.3.2 stable release. The system allows you to configure it, but it hasn't been wired up end-to-end in that specific version.

Prevention

Always check the release notes for your OpenClaw version to understand which models are actively supported in runtime. If you encounter issues with newer models, consider using a development build or pinning to a stable, older model until a patched release is available. Regularly checking the openclaw models status --probe command can help identify missing models before they cause runtime failures.

Last Updated: March 2026

Did this guide solve your problem?