Skip to content

DeepSeek V4 Flash Toolcall Fails with reasoning_content Error #14938

@earthking11

Description

@earthking11

Description

When using deepseek-v4-flash model with Hermes Agent, tool calls fail with the following error:

HTTP 400: The reasoning_content in the thinking mode must be passed back to the API

Root Cause

The deepseek-v4-flash model has thinking mode enabled by default. When the model makes a tool call, the API returns a reasoning_content field in the response. According to DeepSeek's API requirements, subsequent requests must include this reasoning_content to maintain conversation continuity.

However, Hermes currently does not handle the reasoning_content field from DeepSeek API responses, causing the tool call to fail on the next request.

Technical Details

Metadata

Metadata

Assignees

No one assigned

    Labels

    P1High — major feature broken, no workaroundcomp/agentCore agent loop, run_agent.py, prompt builderprovider/deepseekDeepSeek APItype/bugSomething isn't working

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions