Agent Mode not available when using custom model endpoint for Sonnet 3.7

[istranic]

Before submitting your bug report

• [x] I believe this is a bug. I'll try to join the Continue Discord for questions
• [x] I'm not able to find an open issue that reports the same bug
• [x] I've seen the troubleshooting guide on the Continue Docs

Relevant environment info

- OS: Mac
- Continue version: v1.0.11
- IDE version: VS Code 1.95
- config:
  
name: Local Assistant
version: 1.0.0
schema: v1
models:
  - name: Claude 3.7
    provider: openai
    model: anthropic.claude-3-7-sonnet
    apiBase: .....
    apiKey: ....
context:
  - provider: code
  - provider: docs
  - provider: diff
  - provider: terminal
  - provider: problems
  - provider: folder
  - provider: codebase

Description

I'm using Claude 3.7 through a model proxy and Chat and Edit mode work well. However, Agent mode is Grayed out even though the model should work with Agent mode.

I'm assuming this is because the model proxy is not recognized as a model that is compatible with Agent mode. Is there anything I can modify in the config to fix this?

To reproduce

No response

Log output

[harkamals]

--> Claude hosted on Databricks has issues with Agent mode

I am also stuck with this, although I am able to enable Agent mode "capabilities": { "tools": true },

-- but it does not work "Error (status code 400) no body"

please let me know if it works somehow.

--> EDIT: GPT4o hosted on Databricks and agent mode works ok.

[harkamals]

Hi @istranic does EDIT mode work with claude (databricks hosted) ?

It does not for me "INVALID_PARAMETER_VALUE: Missing required Chat parameter: 'messages'"

[istranic]

I'm not using Databricks so I'm not sure. In my case, I was able to get a bit further with agent mode by adding this to the config yaml

    capabilities: [tool_use, image_input]

Now I get some errors from our model endpoint, but I suspect those are not caused by Continue

[wangjl1993]

I encountered the same issue! I am using various models with the AiHubMix node, but Claude and Gemini can only be used in chat mode. Here is my config:

- name: Claude-3.7
    provider: openai
    model: claude-3-7-sonnet-20250219
    apiBase: https://aihubmix.com/v1
    apiKey: xxx
    chatOptions:
      baseSystemMessage: >-
        <important_rules>
          Always include the language and file name in the info string when you write code blocks. 
          If you are editing "src/main.py" for example, your code block should start with '```python src/main.py'

          When addressing code modification requests, present a concise code snippet that
          emphasizes only the necessary changes and uses abbreviated placeholders for
          unmodified sections. For example:

          ```language /path/to/file
          // ... rest of code here ...

          {{ modified code here }}

          // ... rest of code here ...

          {{ another modification }}

          // ... rest of code here ...
          ```

          In existing files, you should always restate the function or class that the snippet belongs to:

          ```language /path/to/file
          // ... rest of code here ...
          
          function exampleFunction() {
            // ... rest of code here ...
            
            {{ modified code here }}
            
            // ... rest of code here ...
          }
          
          // ... rest of code here ...
          ```

          Since users have access to their complete file, they prefer reading only the
          relevant modifications. It's perfectly acceptable to omit unmodified portions
          at the beginning, middle, or end of files using these "lazy" comments. Only
          provide the complete file when explicitly requested. Include a concise explanation
          of changes unless the user specifically asks for code only.
        </important_rules>


        You are an expert software developer. You give helpful and concise
        responses.
    roles:
      - chat
      - edit
      - apply

[Hunter6324]

For the same issue, gemini-2.5-pro and claude-sonnet-4 on OpenAI composite cannot be used in Agent mode, only in Chat mode. GPT-4.1 works in the same way

models:

• name: gpt-4.1 provider: openai model: gpt-4.1 apiBase: http://192.168.50.180:4141/v1 apiKey: _
• name: claude-sonnet-4 provider: openai model: claude-sonnet-4 apiBase: http://192.168.50.180:4141/v1 apiKey: _
• name: gemini-2.5-pro provider: openai model: gemini-2.5-pro apiBase: http://192.168.50.180:4141/v1 apiKey: _

[CooperCorona]

Getting what I believe to be the same issue using Qwen3 and OpenWebUI (which is OpenAI-API-compatible). Running the model locally on my MacBook via Ollama allows for agent mode, but it runs slowly. But the remote version, which connects via OpenWebUI, has agent mode disabled. If I manually add "capabilities": { "tools": true }, I can use agent mode, but it immediately returns with 400 status code (no body) (so quickly that I'm not even sure OpenWebUI is loading the model into memory and producing output).

models:

        {
            "title": "Qwen3:4b",
            "provider": "ollama",
            "model": "qwen3:4b"
        },
        {
            "title": "Qwen3:4b (remote)",
            "provider": "openai",
            "apiKey": "REDACTED",
            "apiBase": "REDACTED",
            "model": "qwen3:4b",
            "useLegacyCompletionsEndpoint": false
        },

[lee-b]

If agent mode only requires tool use, it would be much better if continue checked for that by asking the model to call a tool and verifying that the tool is called, imo.

[github-actions[bot]]

This issue hasn't been updated in 90 days and will be closed after an additional 10 days without activity. If it's still important, please leave a comment and share any new information that would help us address the issue.