silicon custom model manage

[MJ-0414]

Self Checks

• I have read the Contributing Guide and Language Policy.
• This is only for bug report, if you would like to ask a question, please head to Discussions.
• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report, otherwise it will be closed.
• 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：）
• Please do not modify this template :) and fill in all the required fields.

Dify version

V1.11.1

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

When I add custom model credentials in Silicon and the addition is successful, I cannot select the newly added custom model in the LLM node. The new model is “Qwen/Qwen3-VL-32B-Instruct” and "Qwen/Qwen3-VL-235B-A22B-Instruct"

✔️ Expected Behavior

When I can select the newly added custom model in the LLM node of the workflow

❌ Actual Behavior

No response

[dosubot]

This is a known issue in v1.11.1: after adding custom model credentials in Silicon, the new model may not appear in the LLM node selection because the UI state does not refresh properly and/or the model credential is not explicitly added to the active model list. A recent fix (PR langgenius/dify#30250, merged after v1.11.1) ensures that after adding or editing a custom model, the UI and backend state are updated so the model becomes selectable.

In v1.11.1, you may need to manually ensure that after adding credentials, the model is added to the active model list (not left in a pending state). If the UI does not reflect the new model, try refreshing the page or reloading the workflow editor. If the issue persists, upgrading to a version that includes this fix is the most reliable solution.

The backend logic requires that the model credential be explicitly added to the model list for it to become available for selection; otherwise, it will not show up in the LLM node even if credentials were added successfully. This is handled by the add_model_credential_to_model_list method in the backend service. If you have access, you can check the model's status in the backend or database to confirm whether it is active or still pending.

Reference:

• PR langgenius/dify#30250 - fix custom tool content is not update
• ModelProviderService code

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[dosubot]

This is a known issue in Dify v1.11.1: after adding custom model credentials in Silicon, the new model may not appear in the LLM node because the UI and backend state do not always sync properly, and the model credential might not be added to the active model list. A fix for this was merged after v1.11.1 (PR #30250), which ensures that after adding or editing a custom model, the UI and backend state are updated so the model becomes selectable.

For v1.11.1, try these steps:

• Refresh the page or reload the workflow editor after adding a custom model.
• Make sure the model is in the active model list (not pending) in the backend or database.
• Confirm the model supports required capabilities (function call/ReACT) and is fully configured as a system inference model.
• Remove or configure any other model plugins (like Ollama or vLLM) that are unconfigured or unavailable, as these can interfere with model selection.
• Double-check environment variables like POSITION_PROVIDER_INCLUDES and POSITION_PROVIDER_EXCLUDES.
• Restart Docker containers after changes and check backend logs for errors.

If the issue persists, upgrading to a version that includes the fix from PR #30250 is the most reliable solution. For more details, see this issue.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}