Skip to content

fix(reply): clear 'model unavailable' reply + correct operator hint for retired runtime models#97611

Merged
RomneyDa merged 2 commits into
openclaw:mainfrom
RomneyDa:fix/runtime-bound-model-resolution-hint
Jun 29, 2026
Merged

fix(reply): clear 'model unavailable' reply + correct operator hint for retired runtime models#97611
RomneyDa merged 2 commits into
openclaw:mainfrom
RomneyDa:fix/runtime-bound-model-resolution-hint

Conversation

@RomneyDa

@RomneyDa RomneyDa commented Jun 29, 2026 •
edited
Loading

Copy link
Copy Markdown
Member

What Problem This Solves

Fixes an issue where a chat user (e.g. on Telegram) gets "⚠️ Something went
wrong while processing your request. Please try again, or use /new to start a
fresh session" on every message — and /new doesn't help — when the agent's
configured model has been retired/renamed by the provider.

Concrete trigger: a config pinned to openai/gpt-5.3-codex via the codex
agent runtime (OpenAI ChatGPT account). After OpenAI stopped offering
gpt-5.3-codex to ChatGPT-account Codex, every turn failed. Two things made it
hard to recover:

  1. Chat user surface: model resolution throws a typed FailoverError with
    reason model_not_found, but the reply classifier has no matching branch, so
    the reply falls back to the generic "try again / use /new" copy — which is
    actively wrong here, since retrying or starting a new session can never fix a
    bad model id.

  2. Operator surface: the underlying "Unknown model" error appends a hint
    telling the operator to register the model in models.providers[].models[].
    For a runtime-bound model that's misleading: runtime-owned models draw their
    catalog from the runtime/account, not config, so following the hint can't
    help (and can push the operator toward a config that the provider rejects
    downstream).

Why This Change Was Made

Two small, layered changes, one per surface:

  • Chat reply (provider-request-error-classifier.ts): add a
    provider_model_unavailable classification that returns a dedicated reply
    ("the model is unavailable; this needs a config change; retrying//new won't
    help") instead of the generic copy. Detection is structural and typed-only
    — it keys off isFailoverError(err) && err.reason === "model_not_found" (the
    reason run.ts sets at resolution time). It does not match provider error
    text.

  • Operator hint (embedded-agent-runner/model.ts): when the configured
    agents.defaults.models entry is bound to an agentRuntime, stop advising a
    models.providers[] registration and instead point at the runtime's live
    catalog (openclaw models list --provider <id>).

Non-runtime/plain-provider behavior is unchanged.

Scope / non-goals

The chat classifier deliberately handles only the typed model_not_found
failover. It does not classify raw provider error strings or account-scoped
provider 400s (e.g. OpenAI's "model is not supported when using Codex with a
ChatGPT account"). Those carry no structured signal, and string-matching error
classification belongs to the failover layer that already owns it — not to this
downstream reply classifier. In practice the typed path covers the real
recovery case, and the operator-hint change steers operators away from the
config state that produces such provider 400s.

User Impact

  • Chat users now get an accurate, actionable reply instead of an endless
    generic "try again" loop, for any failure that surfaces as a typed
    model_not_found.
  • Operators debugging via logs/CLI get a hint that leads to the real fix
    (switch to an available model) instead of a models.providers[] registration
    that can't help for runtime-bound models.

Evidence

  • New unit tests:
    • provider-request-error-classifier.test.ts — a typed
      FailoverError(reason: "model_not_found") classifies to
      provider_model_unavailable; a bare error string without the typed reason,
      and a different typed reason (overloaded), both stay unclassified
      (locks in structural-only detection).
    • model.test.ts — a runtime-bound entry gets the runtime-catalog hint; a
      plain-provider entry keeps the existing registration hint.
  • Reproduced live on an affected install: the Codex app-server offered
    gpt-5.5/gpt-5.4/gpt-5.4-mini but not gpt-5.3-codex; the failing path
    threw exactly the typed model_not_found failover handled here, and switching
    the default to openai/gpt-5.5 restored normal operation. (Live
    ChatGPT/Codex account path not re-run in CI.)

AI-assisted: investigated and authored with Claude Code.

@openclaw-barnacle openclaw-barnacle Bot added agents Agent runtime and tooling size: XS maintainer Maintainer-authored PR labels Jun 29, 2026
@clawsweeper

clawsweeper Bot commented Jun 29, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Codex review: needs maintainer review before merge. Reviewed June 28, 2026, 9:56 PM ET / 01:56 UTC.

Summary
The branch adds a typed model_not_found auto-reply classification and changes unknown-model hints for agentRuntime-bound agents.defaults.models entries to point at the runtime catalog instead of models.providers[], with focused regression tests.

PR surface: Source +10, Tests +23. Total +33 across 2 files.

Reproducibility: yes. at source level: current main throws typed model_not_found for unresolved models, lacks a matching reply-classifier branch, and emits provider-registration advice for configured agent model rows without provider rows. I did not rerun the live ChatGPT/Codex account path in this read-only review.

Review metrics: none identified.

Merge readiness
Overall: 🐚 platinum hermit
Proof: 🌊 off-meta tidepool
Patch quality: 🐚 platinum hermit
Result: ready for maintainer review.

Overall follows the weaker of proof and patch quality, so missing proof can cap an otherwise strong patch.

Rank-up moves:

  • Align the PR body or release context with the actual typed model_not_found behavior if maintainers want the public description to mention provider 400 text handling.

Risk before merge

  • [P1] The PR body still describes raw provider-message/account-scoped 400 classification, but the latest diff intentionally handles only typed FailoverError reason model_not_found; release notes should describe the typed path unless maintainers add the broader text classifier.

Maintainer options:

  1. Decide the mitigation before merge
    Land the focused classifier and hint change after maintainer review and green checks, with PR body or release context aligned to the actual typed-error scope.
  2. Pause or close
    Do not merge this PR until maintainers decide whether the risk is worth taking.

Next step before merge

  • No automated repair candidate: this is a protected maintainer-labeled PR with no concrete patch defect from review; normal maintainer review and CI should decide landing.

Security
Cleared: Cleared: the diff changes TypeScript error classification/copy and tests only, with no dependency, workflow, secret, package, or supply-chain surface.

Review details

Best possible solution:

Land the focused classifier and hint change after maintainer review and green checks, with PR body or release context aligned to the actual typed-error scope.

Do we have a high-confidence way to reproduce the issue?

Yes at source level: current main throws typed model_not_found for unresolved models, lacks a matching reply-classifier branch, and emits provider-registration advice for configured agent model rows without provider rows. I did not rerun the live ChatGPT/Codex account path in this read-only review.

Is this the best way to solve the issue?

Yes, with one scope note: the typed FailoverError classifier is the right chat-user layer and the hint builder is the narrow operator layer. The current diff should not be treated as raw provider-400 text classification unless that behavior is added or the PR body is corrected.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning high; reviewed against 4a4657a1828f.

Label changes

Label changes:

  • add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
  • remove rating: 🦞 diamond lobster: Current PR rating is rating: 🐚 platinum hermit, so this older rating label is no longer current.

Label justifications:

  • P2: This fixes a real user-facing model-recovery failure with limited blast radius in error classification and recovery copy.
  • rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
  • status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR. Not applicable: Not applicable as a maintainer/MEMBER PR; the PR body reports live affected-install reproduction and the contributor proof gate does not require action here.
Evidence reviewed

PR surface:

Source +10, Tests +23. Total +33 across 2 files.

View PR surface stats
Area Files Added Removed Net
Source 1 14 4 +10
Tests 1 23 0 +23
Docs 0 0 0 0
Config 0 0 0 0
Generated 0 0 0 0
Other 0 0 0 0
Total 2 37 4 +33

What I checked:

  • Live PR state: Live GitHub metadata shows this PR is open, mergeable, authored by a MEMBER, labeled maintainer, and currently at head a436ebf8af7f0f7ce07c7a82f6b7e15d73ff2c4f; cleanup review should keep it open for normal maintainer handling. (a436ebf8af7f)
  • Current-main operator hint: Current main builds Unknown model errors by appending buildMissingProviderModelRegistrationHint before provider plugin hints; the existing helper emits models.providers[] registration advice whenever a configured agents.defaults.models entry lacks a matching provider model row. (src/agents/embedded-agent-runner/model.ts:1879, 4a4657a1828f)
  • PR operator hint fix: At the PR head, buildMissingProviderModelRegistrationHint now detects configuredEntry.agentRuntime?.id and returns runtime-catalog guidance instead of provider-registration guidance for runtime-bound models. (src/agents/embedded-agent-runner/model.ts:1921, a436ebf8af7f)
  • Current-main chat failure path: When model resolution fails, current main throws a typed FailoverError with reason model_not_found; the existing reply classifier has auth, 429, conversation-state, and internal-error branches but no model-unavailable branch. (src/agents/embedded-agent-runner/run.ts:1142, 4a4657a1828f)
  • PR chat classifier fix: At the PR head, the classifier adds provider_model_unavailable and returns the dedicated user message only for typed FailoverError instances whose reason is model_not_found; tests also assert raw text alone remains unclassified. (src/auto-reply/reply/provider-request-error-classifier.ts:61, a436ebf8af7f)
  • Regression coverage: The PR keeps the plain provider-registration hint test, adds a runtime-bound model hint test, and adds classifier tests for typed model_not_found, raw-text non-classification, and other typed failover reasons. (src/agents/embedded-agent-runner/model.test.ts:2811, a436ebf8af7f)

Likely related people:

  • sweetcornna: Authored the merged provider-registration hint change that introduced the current recovery text this PR narrows. (role: recent area contributor; confidence: high; commits: 1532b46e2a63; files: src/agents/embedded-agent-runner/model.ts, src/agents/embedded-agent-runner/model.test.ts)
  • vincentkoc: Merged the provider-registration hint PR and has recent merged work in the central model resolver file. (role: merger and recent agent-runner contributor; confidence: medium; commits: 1532b46e2a63, 206552c6977b, 6ee989a235f0; files: src/agents/embedded-agent-runner/model.ts, src/agents/embedded-agent-runner/model.test.ts)
  • sjf-oa: Recently added typed provider-auth failure handling in the same provider request classifier and test file. (role: recent classifier contributor; confidence: high; commits: 113d6f3c6483; files: src/auto-reply/reply/provider-request-error-classifier.ts, src/auto-reply/reply/provider-request-error-classifier.test.ts)
  • dutifulbob: Introduced the provider conversation-state classifier path that this PR follows for user-facing channel failure copy. (role: classifier feature contributor; confidence: medium; commits: 80d03a1e5bcc; files: src/auto-reply/reply/provider-request-error-classifier.ts, src/auto-reply/reply/provider-request-error-classifier.test.ts)
  • steipete: Recent OpenAI/Codex model catalog and docs work is adjacent to the runtime-catalog contract behind this fix. (role: Codex/OpenAI catalog contributor; confidence: medium; commits: f4b5e58231c7, 33c0cd1378fc, 20b05f220e67; files: extensions/codex/provider.ts, extensions/openai/openai-provider.ts, docs/concepts/model-providers.md)
What the crustacean ranks mean
  • 🦀 challenger crab: rare, exceptional readiness with strong proof, clean implementation, and convincing validation.
  • 🦞 diamond lobster: very strong readiness with only minor maintainer review expected.
  • 🐚 platinum hermit: good normal PR, likely mergeable with ordinary maintainer review.
  • 🦐 gold shrimp: useful signal, but proof or patch confidence is still limited.
  • 🦪 silver shellfish: thin signal; proof, validation, or implementation needs work.
  • 🧂 unranked krab: not merge-ready because proof is missing/unusable or there are serious correctness or safety concerns.
  • 🌊 off-meta tidepool: rating does not apply to this item.

Shiny media proof means a screenshot, video, or linked artifact directly shows the changed behavior. Runtime, network, CSP, and security claims still need visible diagnostics.

How this review workflow works
  • ClawSweeper keeps one durable marker-backed review comment per issue or PR.
  • Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
  • A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
  • PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
  • Maintainers can also comment @clawsweeper review to request a fresh review only.
  • Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge.
  • Maintainer-only repair and merge flows require explicit commands such as @clawsweeper autofix, @clawsweeper automerge, @clawsweeper fix ci, or @clawsweeper address review.
  • Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

@clawsweeper clawsweeper Bot added rating: 🦞 diamond lobster Very strong PR readiness with only minor maintainer review expected. status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR. P2 Normal backlog priority with limited blast radius. labels Jun 29, 2026
@RomneyDa RomneyDa changed the title fix(models): don't advise models.providers[] registration for runtime-bound models fix(reply): clear 'model unavailable' reply + correct operator hint for retired runtime models Jun 29, 2026
RomneyDa and others added 2 commits June 28, 2026 18:52
@RomneyDa @claude
fix(models): don't advise models.providers[] registration for runtime…
586912a 
…-bound models

When an agent's configured model is bound to an agent runtime (e.g. the
"codex" runtime, whose catalog comes from the OpenAI ChatGPT-account
app-server), and that model id is no longer offered by the runtime, model
resolution fails with "Unknown model: <provider>/<model>". The appended hint
told users to register the model in models.providers[].models[].

For runtime-owned models that advice is misleading: adding the registration
makes resolution "succeed" only for the request to be rejected later by the
provider — e.g. OpenAI returns 400 "model is not supported when using Codex
with a ChatGPT account" once a model id such as gpt-5.3-codex is deprecated.

Detect the agentRuntime binding on the configured agents.defaults.models entry
and instead point the user at the runtime's live catalog (openclaw models list
--provider <id>) and at switching the configured default to an available model.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
@RomneyDa @claude
fix(reply): show a clear "model unavailable" reply instead of generic…
a436ebf 
… failure

When a configured model is retired/renamed by the provider, model resolution
fails (run.ts throws a typed FailoverError with reason "model_not_found") and
the agent runner falls back to the generic "Something went wrong … use /new"
reply. That copy is actively misleading for this failure: retrying or starting
a new session can never help, because the model id itself must be changed in
config. Users on a deprecated model (e.g. gpt-5.3-codex on a Codex ChatGPT
account) just see the generic message on every message and on /new.

Classify this failure into a dedicated user-facing reply that explains the
model is unavailable and points at the config. Detection is structural — it
keys off the typed FailoverError reason, not provider error text — so it stays
robust as provider wording changes; free-text rejections without a typed reason
remain the responsibility of the failover layer that owns error classification.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
@RomneyDa
RomneyDa force-pushed the fix/runtime-bound-model-resolution-hint branch from 614ac49 to a436ebf Compare June 29, 2026 01:52
@clawsweeper clawsweeper Bot added rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. and removed rating: 🦞 diamond lobster Very strong PR readiness with only minor maintainer review expected. labels Jun 29, 2026
@RomneyDa
RomneyDa merged commit 9388550 into openclaw:main Jun 29, 2026
115 of 117 checks passed
@github-actions github-actions Bot mentioned this pull request Jun 29, 2026
Open
github-actions Bot pushed a commit to Desicool/openclaw that referenced this pull request Jun 29, 2026
@RomneyDa @claude
fix(reply): clear 'model unavailable' reply + correct operator hint f…
e819ec3 
…or retired runtime models (openclaw#97611)

* fix(models): don't advise models.providers[] registration for runtime-bound models

When an agent's configured model is bound to an agent runtime (e.g. the
"codex" runtime, whose catalog comes from the OpenAI ChatGPT-account
app-server), and that model id is no longer offered by the runtime, model
resolution fails with "Unknown model: <provider>/<model>". The appended hint
told users to register the model in models.providers[].models[].

For runtime-owned models that advice is misleading: adding the registration
makes resolution "succeed" only for the request to be rejected later by the
provider — e.g. OpenAI returns 400 "model is not supported when using Codex
with a ChatGPT account" once a model id such as gpt-5.3-codex is deprecated.

Detect the agentRuntime binding on the configured agents.defaults.models entry
and instead point the user at the runtime's live catalog (openclaw models list
--provider <id>) and at switching the configured default to an available model.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

* fix(reply): show a clear "model unavailable" reply instead of generic failure

When a configured model is retired/renamed by the provider, model resolution
fails (run.ts throws a typed FailoverError with reason "model_not_found") and
the agent runner falls back to the generic "Something went wrong … use /new"
reply. That copy is actively misleading for this failure: retrying or starting
a new session can never help, because the model id itself must be changed in
config. Users on a deprecated model (e.g. gpt-5.3-codex on a Codex ChatGPT
account) just see the generic message on every message and on /new.

Classify this failure into a dedicated user-facing reply that explains the
model is unavailable and points at the config. Detection is structural — it
keys off the typed FailoverError reason, not provider error text — so it stays
robust as provider wording changes; free-text rejections without a typed reason
remain the responsibility of the failover layer that owns error classification.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <[email protected]>
@clawsweeper clawsweeper Bot mentioned this pull request Jun 29, 2026
Closed
QiuYuang pushed a commit to QiuYuang/openclaw that referenced this pull request Jul 1, 2026
@RomneyDa @claude @QiuYuang
fix(reply): clear 'model unavailable' reply + correct operator hint f…
980e6e8 
…or retired runtime models (openclaw#97611)

* fix(models): don't advise models.providers[] registration for runtime-bound models

When an agent's configured model is bound to an agent runtime (e.g. the
"codex" runtime, whose catalog comes from the OpenAI ChatGPT-account
app-server), and that model id is no longer offered by the runtime, model
resolution fails with "Unknown model: <provider>/<model>". The appended hint
told users to register the model in models.providers[].models[].

For runtime-owned models that advice is misleading: adding the registration
makes resolution "succeed" only for the request to be rejected later by the
provider — e.g. OpenAI returns 400 "model is not supported when using Codex
with a ChatGPT account" once a model id such as gpt-5.3-codex is deprecated.

Detect the agentRuntime binding on the configured agents.defaults.models entry
and instead point the user at the runtime's live catalog (openclaw models list
--provider <id>) and at switching the configured default to an available model.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

* fix(reply): show a clear "model unavailable" reply instead of generic failure

When a configured model is retired/renamed by the provider, model resolution
fails (run.ts throws a typed FailoverError with reason "model_not_found") and
the agent runner falls back to the generic "Something went wrong … use /new"
reply. That copy is actively misleading for this failure: retrying or starting
a new session can never help, because the model id itself must be changed in
config. Users on a deprecated model (e.g. gpt-5.3-codex on a Codex ChatGPT
account) just see the generic message on every message and on /new.

Classify this failure into a dedicated user-facing reply that explains the
model is unavailable and points at the config. Detection is structural — it
keys off the typed FailoverError reason, not provider error text — so it stays
robust as provider wording changes; free-text rejections without a typed reason
remain the responsibility of the failover layer that owns error classification.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <[email protected]>
chenyangjun-xy pushed a commit to chenyangjun-xy/openclaw that referenced this pull request Jul 1, 2026
@RomneyDa @claude chenyangjun-xy
fix(reply): clear 'model unavailable' reply + correct operator hint f…
782d42f 
…or retired runtime models (openclaw#97611)

* fix(models): don't advise models.providers[] registration for runtime-bound models

When an agent's configured model is bound to an agent runtime (e.g. the
"codex" runtime, whose catalog comes from the OpenAI ChatGPT-account
app-server), and that model id is no longer offered by the runtime, model
resolution fails with "Unknown model: <provider>/<model>". The appended hint
told users to register the model in models.providers[].models[].

For runtime-owned models that advice is misleading: adding the registration
makes resolution "succeed" only for the request to be rejected later by the
provider — e.g. OpenAI returns 400 "model is not supported when using Codex
with a ChatGPT account" once a model id such as gpt-5.3-codex is deprecated.

Detect the agentRuntime binding on the configured agents.defaults.models entry
and instead point the user at the runtime's live catalog (openclaw models list
--provider <id>) and at switching the configured default to an available model.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

* fix(reply): show a clear "model unavailable" reply instead of generic failure

When a configured model is retired/renamed by the provider, model resolution
fails (run.ts throws a typed FailoverError with reason "model_not_found") and
the agent runner falls back to the generic "Something went wrong … use /new"
reply. That copy is actively misleading for this failure: retrying or starting
a new session can never help, because the model id itself must be changed in
config. Users on a deprecated model (e.g. gpt-5.3-codex on a Codex ChatGPT
account) just see the generic message on every message and on /new.

Classify this failure into a dedicated user-facing reply that explains the
model is unavailable and points at the config. Detection is structural — it
keys off the typed FailoverError reason, not provider error text — so it stays
robust as provider wording changes; free-text rejections without a typed reason
remain the responsibility of the failover layer that owns error classification.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <[email protected]>
chenyangjun-xy pushed a commit to chenyangjun-xy/openclaw that referenced this pull request Jul 3, 2026
@RomneyDa @claude @chenyangjun-xy
fix(reply): clear 'model unavailable' reply + correct operator hint f…
0dee97a 
…or retired runtime models (openclaw#97611)

* fix(models): don't advise models.providers[] registration for runtime-bound models

When an agent's configured model is bound to an agent runtime (e.g. the
"codex" runtime, whose catalog comes from the OpenAI ChatGPT-account
app-server), and that model id is no longer offered by the runtime, model
resolution fails with "Unknown model: <provider>/<model>". The appended hint
told users to register the model in models.providers[].models[].

For runtime-owned models that advice is misleading: adding the registration
makes resolution "succeed" only for the request to be rejected later by the
provider — e.g. OpenAI returns 400 "model is not supported when using Codex
with a ChatGPT account" once a model id such as gpt-5.3-codex is deprecated.

Detect the agentRuntime binding on the configured agents.defaults.models entry
and instead point the user at the runtime's live catalog (openclaw models list
--provider <id>) and at switching the configured default to an available model.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

* fix(reply): show a clear "model unavailable" reply instead of generic failure

When a configured model is retired/renamed by the provider, model resolution
fails (run.ts throws a typed FailoverError with reason "model_not_found") and
the agent runner falls back to the generic "Something went wrong … use /new"
reply. That copy is actively misleading for this failure: retrying or starting
a new session can never help, because the model id itself must be changed in
config. Users on a deprecated model (e.g. gpt-5.3-codex on a Codex ChatGPT
account) just see the generic message on every message and on /new.

Classify this failure into a dedicated user-facing reply that explains the
model is unavailable and points at the config. Detection is structural — it
keys off the typed FailoverError reason, not provider error text — so it stays
robust as provider wording changes; free-text rejections without a typed reason
remain the responsibility of the failover layer that owns error classification.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

---------

Co-authored-by: Claude Opus 4.8 (1M context) <[email protected]>
This was referenced Jul 4, 2026
Closed
Closed
Closed
@ObliviateRickLin ObliviateRickLin mentioned this pull request Jul 12, 2026
Merged
ObliviateRickLin added a commit to ObliviateRickLin/openclaw that referenced this pull request Jul 12, 2026
@ObliviateRickLin
fix(agents): classify account-restricted model 400s as model_not_found
6933e9d 
OpenAI-backed runtimes reject plan/account-restricted models with HTTP
400 invalid_request_error ("The '<model>' model is not supported when
using Codex with a ChatGPT account."). Without a model_not_found match
the ambiguous-400 branch collapses this into a format failure, so users
get the generic retry//new copy for a config-only problem (openclaw#104490).
The 'when using' qualifier keeps capability rejections ("not supported
for tool calling") out of the class, preserving the openclaw#97611 contract.

Fixes openclaw#104490
ObliviateRickLin added a commit to ObliviateRickLin/openclaw that referenced this pull request Jul 12, 2026
@ObliviateRickLin
fix(agents): classify account-restricted model 400s as model_not_found
7abbb6f 
OpenAI-backed runtimes reject plan/account-restricted models with HTTP
400 invalid_request_error ("The '<model>' model is not supported when
using Codex with a ChatGPT account."). Without a model_not_found match
the ambiguous-400 branch collapses this into a format failure, so users
get the generic retry//new copy for a config-only problem (openclaw#104490).
The 'when using' qualifier keeps capability rejections ("not supported
for tool calling") out of the class, preserving the openclaw#97611 contract.

Fixes openclaw#104490
altaywtf added a commit that referenced this pull request Jul 13, 2026
@ObliviateRickLin @altaywtf
fix(agents): classify account-restricted model 400s as model_not_found (
21c3778 
#104878)

* fix(agents): classify account-restricted model 400s as model_not_found

OpenAI-backed runtimes reject plan/account-restricted models with HTTP
400 invalid_request_error ("The '<model>' model is not supported when
using Codex with a ChatGPT account."). Without a model_not_found match
the ambiguous-400 branch collapses this into a format failure, so users
get the generic retry//new copy for a config-only problem (#104490).
The 'when using' qualifier keeps capability rejections ("not supported
for tool calling") out of the class, preserving the #97611 contract.

Fixes #104490

* fix(agents): narrow account model rejection match

---------

Co-authored-by: Altay <[email protected]>
wm0018 pushed a commit to wm0018/openclaw that referenced this pull request Jul 14, 2026
@ObliviateRickLin @altaywtf @wm0018
fix(agents): classify account-restricted model 400s as model_not_found (
5f6269b 
openclaw#104878)

* fix(agents): classify account-restricted model 400s as model_not_found

OpenAI-backed runtimes reject plan/account-restricted models with HTTP
400 invalid_request_error ("The '<model>' model is not supported when
using Codex with a ChatGPT account."). Without a model_not_found match
the ambiguous-400 branch collapses this into a format failure, so users
get the generic retry//new copy for a config-only problem (openclaw#104490).
The 'when using' qualifier keeps capability rejections ("not supported
for tool calling") out of the class, preserving the openclaw#97611 contract.

Fixes openclaw#104490

* fix(agents): narrow account model rejection match

---------

Co-authored-by: Altay <[email protected]>
github-actions Bot pushed a commit to Desicool/openclaw that referenced this pull request Jul 14, 2026
@ObliviateRickLin @altaywtf
fix(agents): classify account-restricted model 400s as model_not_found (
7020984 
openclaw#104878)

* fix(agents): classify account-restricted model 400s as model_not_found

OpenAI-backed runtimes reject plan/account-restricted models with HTTP
400 invalid_request_error ("The '<model>' model is not supported when
using Codex with a ChatGPT account."). Without a model_not_found match
the ambiguous-400 branch collapses this into a format failure, so users
get the generic retry//new copy for a config-only problem (openclaw#104490).
The 'when using' qualifier keeps capability rejections ("not supported
for tool calling") out of the class, preserving the openclaw#97611 contract.

Fixes openclaw#104490

* fix(agents): narrow account model rejection match

---------

Co-authored-by: Altay <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

agents Agent runtime and tooling maintainer Maintainer-authored PR P2 Normal backlog priority with limited blast radius. rating: 🐚 platinum hermit Good normal PR readiness with ordinary maintainer review expected. size: S status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR.

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@RomneyDa