Description
I'm having an issue using Claude Opus 5 and Claude Sonnet 5 with Agent Zero through my own OpenAI-compatible API endpoint.
The API request reaches my API successfully and the requests are being processed/billed, but Agent Zero receives an empty/unusable response.
Interestingly, the exact same Agent Zero setup and API endpoint works correctly with GPT-5.6.
Environment
- Agent Zero: [version]
- OS: Ubuntu 24.04
- Python: 3.12
- LiteLLM: [version]
- Installation: Docker
- API: Custom OpenAI-compatible API
- API mode: OpenAI-compatible / Chat Completions
Models tested
Claude Opus 5
Model:
claude-opus-5
Agent Zero repeatedly shows:
A0: Calling LLM...
WRN: Model returned an empty response (no reasoning, no content, no tool call)
Description
I'm having an issue using Claude Opus 5 and Claude Sonnet 5 with Agent Zero through my own OpenAI-compatible API endpoint.
The API request reaches my API successfully and the requests are being processed/billed, but Agent Zero receives an empty/unusable response.
Interestingly, the exact same Agent Zero setup and API endpoint works correctly with GPT-5.6.
Environment
Models tested
Claude Opus 5
Model:
claude-opus-5Agent Zero repeatedly shows: