Issue Station
Sign In
Tools hereToolbox

openclaw - 💡(How to fix) Fix [Bug]: openai-codex catalog exposes only gpt-5.4 in models list and native /models on 2026.3.13 [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#53755•Fetched 2026-04-08 01:23:52
View on GitHub
Comments
0
Participants
1
Timeline
0
Reactions
0
Author
mastash3ffmastash3ff
Participants
mastash3ffmastash3ff

On OpenClaw 2026.3.13, the openai-codex catalog is exposed as exactly one model (openai-codex/gpt-5.4) in both openclaw models list and Discord native /models, while other providers (for example anthropic) expose multiple models normally.

Root Cause

On OpenClaw 2026.3.13, the openai-codex catalog is exposed as exactly one model (openai-codex/gpt-5.4) in both openclaw models list and Discord native /models, while other providers (for example anthropic) expose multiple models normally.

Code Example

$ openclaw --version
2026.3.13

$ openclaw models list
Model                                      Input      Ctx      Local Auth  Tags
openai-codex/gpt-5.4                       text+image 266k     no    yes   default,configured,alias:code
anthropic/claude-sonnet-4-6                text       195k     no    yes   fallback#1,configured,alias:best
openai/gpt-5-mini                          text       195k     no    yes   fallback#2,configured,alias:balanced
google/gemini-3-flash-preview              text       977k     no    yes   fallback#3,configured,alias:research
google/gemini-3.1-flash-image-preview      text+image 977k     no    yes   image,configured
openai/gpt-5.4                             text       195k     no    yes   configured,alias:quality
openai/gpt-5-nano                          text       195k     no    yes   configured,alias:cheap
anthropic/claude-opus-4-6                  text       195k     no    yes   configured,alias:opus
anthropic/claude-haiku-4-5                 text       195k     no    yes   configured,alias:haiku
openai/gpt-5.4-pro                         text+image 1025k    no    yes   configured

$ openclaw models list | rg '^openai-codex/'
openai-codex/gpt-5.4                       text+image 266k     no    yes   default,configured,alias:code

---

/models anthropic
Models (anthropic · 🔑 token (anthropic:default)) — showing 1-3 of 3 (page 1/1)
- anthropic/claude-haiku-4-5
- anthropic/claude-opus-4-6
- anthropic/claude-sonnet-4-6
Switch: <provider/model>
All: /models anthropic all

---

/models openai-codex all
Models (openai-codex · 🔑 oauth (openai-codex:default)) — showing 1-1 of 1 (page 1/1)
- openai-codex/gpt-5.4
Switch: <provider/model>
RAW_BUFFERClick to expand / collapse

Bug type

Behavior bug (incorrect output/state without crash)

Summary

On OpenClaw 2026.3.13, the openai-codex catalog is exposed as exactly one model (openai-codex/gpt-5.4) in both openclaw models list and Discord native /models, while other providers (for example anthropic) expose multiple models normally.

Steps to reproduce

  1. Start OpenClaw 2026.3.13 with a valid openai-codex:default OAuth profile.
  2. Run openclaw models list.
  3. Observe that the only openai-codex/* entry listed is openai-codex/gpt-5.4.
  4. In Discord, run /models openai-codex all.
  5. Observe that the native command shows 1-1 of 1 and only openai-codex/gpt-5.4.
  6. In Discord, run /models anthropic.
  7. Observe that anthropic shows multiple models (claude-haiku-4-5, claude-opus-4-6, claude-sonnet-4-6) under the same command surface.

Expected behavior

If multiple openai-codex models are supported on this release, openclaw models list and native /models should expose them similarly to other providers instead of collapsing the provider catalog to a single model. If only one openai-codex model is intentionally supported on this release, the docs and model-selection UX should make that explicit.

Actual behavior

openai-codex is exposed as exactly one model in both CLI and Discord native model selection:

  • openclaw models list shows only openai-codex/gpt-5.4
  • Discord /models openai-codex all shows 1-1 of 1
  • Restarting the gateway does not change the result
  • Other providers such as anthropic expose multiple models normally

OpenClaw version

2026.3.13

Operating system

Ubuntu 24.04 / Linux 6.12.10-76061203-generic (x86_64)

Install method

npm global

Model

openai-codex/gpt-5.4

Provider / routing chain

openclaw -> openai-codex (OAuth profile openai-codex:default)

Additional provider/model setup details

Default model is openai-codex/gpt-5.4. Observed allowed model set from openclaw models status --json includes:

  • openai-codex/gpt-5.4
  • anthropic/claude-sonnet-4-6
  • anthropic/claude-opus-4-6
  • anthropic/claude-haiku-4-5
  • openai/gpt-5.4
  • openai/gpt-5-mini
  • openai/gpt-5-nano
  • google/gemini-3-flash-preview
  • google/gemini-3.1-flash-image-preview
  • openai/gpt-5.4-pro

Relevant auth state:

  • openai-codex:default=OAuth
  • anthropic:default=token

This report is specifically about provider catalog exposure/listing behavior for openai-codex, not about Discord session state. Restarting the gateway did not change the result.

Logs, screenshots, and evidence

$ openclaw --version
2026.3.13

$ openclaw models list
Model                                      Input      Ctx      Local Auth  Tags
openai-codex/gpt-5.4                       text+image 266k     no    yes   default,configured,alias:code
anthropic/claude-sonnet-4-6                text       195k     no    yes   fallback#1,configured,alias:best
openai/gpt-5-mini                          text       195k     no    yes   fallback#2,configured,alias:balanced
google/gemini-3-flash-preview              text       977k     no    yes   fallback#3,configured,alias:research
google/gemini-3.1-flash-image-preview      text+image 977k     no    yes   image,configured
openai/gpt-5.4                             text       195k     no    yes   configured,alias:quality
openai/gpt-5-nano                          text       195k     no    yes   configured,alias:cheap
anthropic/claude-opus-4-6                  text       195k     no    yes   configured,alias:opus
anthropic/claude-haiku-4-5                 text       195k     no    yes   configured,alias:haiku
openai/gpt-5.4-pro                         text+image 1025k    no    yes   configured

$ openclaw models list | rg '^openai-codex/'
openai-codex/gpt-5.4                       text+image 266k     no    yes   default,configured,alias:code

Discord native command results:

/models anthropic
Models (anthropic · 🔑 token (anthropic:default)) — showing 1-3 of 3 (page 1/1)
- anthropic/claude-haiku-4-5
- anthropic/claude-opus-4-6
- anthropic/claude-sonnet-4-6
Switch: <provider/model>
All: /models anthropic all
/models openai-codex all
Models (openai-codex · 🔑 oauth (openai-codex:default)) — showing 1-1 of 1 (page 1/1)
- openai-codex/gpt-5.4
Switch: <provider/model>

Gateway restart was attempted and did not change behavior.

Impact and severity

Affected: users relying on openai-codex model selection via CLI or native /models Severity: Medium Frequency: 100% in observed checks on this install Consequence: openai-codex appears to have only one selectable model, which may be either an incorrect catalog exposure bug or undocumented intentional behavior. It makes provider model selection inconsistent across providers and makes it unclear whether additional Codex models are unsupported or just hidden.

Additional information

Closest related issues found were adjacent /model and openai-codex issues, but not this exact symptom. This may be either:

  • a real catalog/listing bug for openai-codex, or
  • expected behavior on 2026.3.13 that is not clearly documented in the model-selection UX/docs.

If maintainers expect only one openai-codex model on this build, a docs/UX clarification would still help.

extent analysis

Fix Plan

To fix the issue of openai-codex being exposed as only one model, we need to update the model catalog exposure logic.

Here are the steps:

  • Update the openclaw code to fetch the list of supported models from the openai-codex provider.
  • Modify the openclaw models list command to display all supported models for each provider.
  • Update the Discord native /models command to display all supported models for each provider.

Example code changes:

# Fetch list of supported models from openai-codex provider
def get_supported_models(provider):
    if provider == 'openai-codex':
        # Fetch list of supported models from openai-codex API
        models = fetch_models_from_openai_codex_api()
        return models
    else:
        # Fetch list of supported models from other providers
        models = fetch_models_from_other_providers(provider)
        return models

# Update openclaw models list command to display all supported models
def list_models():
    providers = get_providers()
    for provider in providers:
        models = get_supported_models(provider)
        for model in models:
            print(f"{provider}/{model}")

# Update Discord native /models command to display all supported models
def handle_models_command(provider):
    models = get_supported_models(provider)
    for model in models:
        print(f"- {provider}/{model}")

Verification

To verify that the fix worked, run the following commands:

  • openclaw models list to check if all supported models are displayed for each provider.
  • /models openai-codex all to check if all supported models are displayed for the openai-codex provider.

Extra Tips

  • Make sure to update the documentation to reflect the changes in the model catalog exposure logic.
  • Test the changes thoroughly to ensure that all supported models are displayed correctly for each provider.
  • Consider adding a feature to allow users to select which models to display for each provider.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

FAQ

Expected behavior

If multiple openai-codex models are supported on this release, openclaw models list and native /models should expose them similarly to other providers instead of collapsing the provider catalog to a single model. If only one openai-codex model is intentionally supported on this release, the docs and model-selection UX should make that explicit.

#conversation history#tool integration#LLM response#prompt template#agent execution

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING