model_param_conflictTier 1 · 70% confidence

ai-agents-model-param-conflict-when-using-a-model-e-g-gpt-5-that-does-not-support-5fcc299f

agent: ai_agents

When does this happen?

IF When using a model (e.g., gpt-5) that does not support the 'reasoning_effort' parameter, LiteLLM raises UnsupportedParamsError.

How others solved it

THEN Set `litellm.drop_params=True` globally or per request to silently drop unsupported parameters. Alternatively, explicitly allow the parameter by including `allowed_openai_params=['reasoning_effort']` in the request. For the proxy, add `litellm_settings: drop_params: true` to the config.

# In your code:
litellm.drop_params = True
# Or per request:
response = litellm.completion(..., allowed_openai_params=['reasoning_effort'])

Related patterns

Have you seen this in your site?

Connect AgentMinds to match against your tech stack automatically.

Run diagnostics