🎚️ feat: Filter OpenRouter Reasoning Efforts by the Selected Model

[berry-13]

Summary

OpenRouter endpoints show a reasoning effort control without extra config. That control, and the "Reasoning Effort" setting in the parameters panels, offered one generic list (none, minimal, low, medium, high, xhigh, max) for every OpenRouter model. OpenRouter rejects an effort a model does not support, so picking none on a mandatory-reasoning model, or max on a model without it, failed at the provider.

OpenRouter reports the efforts each model accepts (reasoning.supported_efforts, reasoning.mandatory) in GET /models. The server now reads that catalog for the OpenRouter endpoints in the app config, caches it for an hour, and serves it at GET /api/endpoints/reasoning-capabilities. The composer selector, the conversation and agent parameter panels and the preset settings narrow the effort list to what the selected model accepts, drop none when reasoning is mandatory, and hide the control for a model the catalog lists without reasoning. The server re-checks a submitted override against the same data, so a stale or hand-built request is refused too.

While the catalog is loading or cannot be fetched, the generic list stays, so a provider outage never removes a control that worked before. Only an endpoint whose base URL host is OpenRouter is fetched, with the administrator's key and never a user's; user-provided keys and URLs are skipped.

How it works

GET {openrouter}/models   (admin key, cached 1h by baseURL+key)
  loadReasoningCapabilities   packages/api/src/endpoints/reasoning.ts
    -> { endpoint: { model: { efforts, mandatory } } }
  GET /api/endpoints/reasoning-capabilities      client queries it for OpenRouter endpoints only
  buildEndpointOption                            loads it only when a request carries an override

applyModelAwareDefaults(..., modelReasoning)     one narrowing step shared by every surface
  undefined  capabilities unknown  -> generic list
  null       model has no reasoning -> control hidden
  { efforts, mandatory }           -> listed efforts (+ Auto), minus none when mandatory

Type of change

• Feature

Testing

Tested environments/configuration:

Jest only. Not run against a live OpenRouter request or in a browser.

Automated tests:

New cases cover the narrowing (supported efforts, mandatory, unknown, no reasoning, unrecognized levels, non-OpenRouter endpoints, explicit definitions untouched), the model variant lookup, the catalog loader (parsing, caching, one fetch per shared key, failed fetch not cached, non-OpenRouter hosts and user-provided URLs skipped), server override validation, lazy capability loading, the client hook and the composer. The narrowing, server rejection, lazy loading and composer cases each failed before the change. Passing: parameterSettings.spec.ts, reasoning.spec.ts, reasoningOverride.spec.ts, custom/config.spec.ts, plus the related client suites (Reasoning, Thinking, ModelPanel, Panel, parameters, useModelReasoning, useChatFunctions.regenerate). lc lint is clean and tsc --noEmit passes in packages/data-provider, packages/api and client.

Screenshots / recordings

Not captured; the changed lists were not checked in a running app.

Risk / compatibility

Adds one outbound request: the server fetches OpenRouter's model catalog at most once an hour per base URL and key, only when an OpenRouter endpoint exists and a session on one asks for it. A model absent from the catalog loses its effort control. buildEndpointOption.js and the endpoints route gain wiring only, with the behavior in packages/api. The buildEndpointOption middleware spec could not run in this environment because @librechat/agents lacks GitHubCompareToolDefinition (unrelated to this change), so that wiring is covered by the applyRequestReasoningOverride tests instead.

Checklist

• I reviewed my own changes
• Relevant tests have been added or updated
• Existing relevant tests pass
• The change does not introduce new warnings or errors
• Required dependency changes have been merged/published
• Required documentation PR: N/A

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-06T21:47:09.059918Z	8d00900	New commits
🔒 Security Review	✅ Completed 2026-10-06T07:33:54.099755Z	450bd3b	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 450bd3b05a

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 439da5998e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: cf9b9de933

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f14c0f3a9f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 612bf16bc8

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d3d4dc1052

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e008deed3f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 31bd79f23c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6abe1c23fb

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d26f2b4ef7

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: a53507c812

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[github-actions]

Lighthouse CI failed. The last 80 log lines contain the measured budgets and assertion failures.

│ 22      │ 'http://localhost:3080/api/permissions/mcpServer/effective/all'                                                 │ 2810.215999999986  │ 3569.832000000024  │ 200    │
│ 23      │ 'http://localhost:3080/api/prompts/groups?limit=10'                                                             │ 2811.8430000000226 │ 4081.8640000000305 │ 200    │
│ 24      │ 'http://localhost:3080/api/keys?name=openAI'                                                                    │ 3168.451000000001  │ 3676.8640000000305 │ 200    │
│ 25      │ 'http://localhost:3080/api/presets'                                                                             │ 3169.8949999999895 │ 3708.0100000000093 │ 200    │
│ 26      │ 'http://localhost:3080/api/tags'                                                                                │ 3171.3239999999932 │ 3713.345000000001  │ 200    │
│ 27      │ 'http://localhost:3080/api/share/link/16390000-0000-4000-8000-000000000001'                                     │ 3171.5830000000133 │ 4082.110000000015  │ 200    │
│ 28      │ 'http://localhost:3080/api/messages/16390000-0000-4000-8000-000000000001?toolPreviews=1'                        │ 3171.8040000000037 │ 4185.582000000024  │ 200    │
│ 29      │ 'http://localhost:3080/api/files/config'                                                                        │ 3173.5130000000063 │ 3962.6380000000063 │ 200    │
│ 30      │ 'http://localhost:3080/api/user/settings/favorites/tools'                                                       │ 3173.7460000000137 │ 4221.078999999998  │ 200    │
│ 31      │ 'http://localhost:3080/api/endpoints/token-config'                                                              │ 3173.9320000000007 │ 4221.331000000035  │ 200    │
│ 32      │ 'http://localhost:3080/api/user/settings/skills/active'                                                         │ 3174.719000000012  │ 4592.937999999995  │ 200    │
│ 33      │ 'http://localhost:3080/api/agents/tools/web_search/auth'                                                        │ 3175.2940000000235 │ 7114.002999999997  │ 200    │
│ 34      │ 'http://localhost:3080/api/agents/tools/calls?conversationId=16390000-0000-4000-8000-000000000001'              │ 3175.8430000000226 │ 4593.928000000014  │ 200    │
│ 35      │ 'http://localhost:3080/api/agents/chat/status/16390000-0000-4000-8000-000000000001?generationProtocolVersion=2' │ 4314.809000000008  │ 4569.786000000022  │ 200    │
└─────────┴─────────────────────────────────────────────────────────────────────────────────────────────────────────────────┴────────────────────┴────────────────────┴────────┘

Inspect .lighthouse HTML/JSON and e2e/lighthouse/README.md. Reuse loaded user/config data; overlap independent reads without bypassing authorization.

┌─────────┬────────────────────────────┬──────────────────────┬───────┐
│ (index) │ audit                      │ median               │ limit │
├─────────┼────────────────────────────┼──────────────────────┼───────┤
│ 0       │ 'largest-contentful-paint' │ 4265.005             │ 4500  │
│ 1       │ 'cumulative-layout-shift'  │ 0.017915458366167585 │ 0.1   │
│ 2       │ 'total-blocking-time'      │ 503.3829999999998    │ 500   │
└─────────┴────────────────────────────┴──────────────────────┴───────┘

  1) [chrome] › e2e/lighthouse/load.spec.ts:10:5 › serial database latency stays within web-vitals budgets 

    Error: Median total-blocking-time must stay within 500

    expect(received).toBeLessThanOrEqual(expected)

    Expected: <= 500
    Received:    503.3829999999998

       at audit.ts:159

      157 |   console.table(measured);
      158 |   for (const { audit, median, limit } of measured) {
    > 159 |     expect(median, `Median ${audit} must stay within ${limit}`).toBeLessThanOrEqual(limit);
          |                                                                 ^
      160 |   }
      161 |   return results;
      162 | }
        at auditPage (/home/runner/work/LibreChat/LibreChat/e2e/lighthouse/audit.ts:159:65)
        at /home/runner/work/LibreChat/LibreChat/e2e/lighthouse/load.spec.ts:33:19

    attachment #1: screenshot (image/png) ──────────────────────────────────────────────────────────
    e2e/lighthouse/.test-results/load-serial-database-latency-stays-within-web-vitals-budgets-chrome/test-failed-1.png
    ────────────────────────────────────────────────────────────────────────────────────────────────

    Error Context: e2e/lighthouse/.test-results/load-serial-database-latency-stays-within-web-vitals-budgets-chrome/error-context.md

    attachment #3: trace (application/zip) ─────────────────────────────────────────────────────────
    e2e/lighthouse/.test-results/load-serial-database-latency-stays-within-web-vitals-budgets-chrome/trace.zip
    Usage:

        npx playwright show-trace e2e/lighthouse/.test-results/load-serial-database-latency-stays-within-web-vitals-budgets-chrome/trace.zip

    ────────────────────────────────────────────────────────────────────────────────────────────────


🤖: global teardown has been started
2026-10-06 18:55:12 �[32minfo�[39m: �[32mMongo Connection options�[39m
2026-10-06 18:55:12 �[32minfo�[39m: �[32m{�[39m
�[32m  "bufferCommands": false�[39m
�[32m}�[39m
🤖:  ✅  Connected to Database
🤖:  ✅  Found user in Database
🤖:  ✅  Deleted 1 convos & 2 messages
🤖:  ✅  Deleted user from Database
🤖: global teardown has been started
2026-10-06 18:55:13 �[32minfo�[39m: �[32mMongo Connection options�[39m
2026-10-06 18:55:13 �[32minfo�[39m: �[32m{�[39m
�[32m  "bufferCommands": false�[39m
�[32m}�[39m
🤖:  ✅  Connected to Database
🤖:  ⚠️  User not found in Database
  1 failed
    [chrome] › e2e/lighthouse/load.spec.ts:10:5 › serial database latency stays within web-vitals budgets 

Open the full run

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c92a276815

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 2ef28057c7

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9fedc9bb92

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 5e0ccf24c2

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 65c64246dd

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 304d9f854d

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".