bug: BadRequestError as gpt-5 only supports temperature=1

[kevin-yauris]

Did you check the docs?

• I have read all the NeMo-Guardrails docs

Is your feature request related to a problem? Please describe.

Hi I am trying to use the latest model for the guardrail I assume better LLM means better performance.
I am trying to use the openai engine with gpt5-mini

DEFAULT_CONFIG_DICT = {"models": [{"type": "main", "engine": "openai", "model": "gpt-5-mini"}]}

but I keep getting error:

openai.BadRequestError: Error code: 400 - {'error': {'message': "Unsupported value:                  
                             'temperature' does not support 0.001 with this model. Only the default (1) value is                  
                             supported.", 'type': 'invalid_request_error', 'param': 'temperature', 'code':                        
                             'unsupported_value

Since I cannot make it works using the built-in engine I try to use a custom engine that calling the needed model behind but the performance is really slow (in average 30 seconds), compared to using the build in engine(under 5 seconds).

Is there any way to use GPT-5 mini or any other latest model with Nemo Guardrail built-in engine?
If it is not, is the custom engine expected to be slow? How to optimize it?

Thank you.

Describe the solution you'd like

To be able to use GPT5-mini without customization

Describe alternatives you've considered

Custom model engine

Additional context

[Pouyanpi]

Hi @kevin-yauris , before we add the fix, below is a workaround:

lowest_temperature: 1

add this to the end of your config file.

for example:

config.yml

models:
  - type: main
    engine: openai
    model: gpt-5-mini
lowest_temperature: 1

Let me know if it unblocked you.

[notetiene]

@Pouyanpi Colang v2 doesn't look to work with gpt5+ models. The reason being that the stop parameter isn't supported.

Here's the resquest response:

LLMCallException: LLM Call Exception: Error code: 400 - {'error': {'message': "Unsupported parameter:
'stop' is not supported with this model.", 'type': 'invalid_request_error', 'param': 'stop', 'code':
'unsupported_parameter'}}