Problems With Eden Chat Response Module

Questions
,
1

:bullseye: What is your goal?

Use the Eden Module to generate a chat response.

:thinking: What is the problem & what have you tried?

Almost everytime in the Eden Generate Chat Response module the first model fails and it has to use the second (Fallback) model.

But last few days I noticed both are failing.

In the error screenshot you can see the first model in this scenario run had some issues with model. But in the module settings we (the users) can not set the specific model to use, we can just select ‘Anthropic’.

So I need to know how to fix this, so I dont waste my credits. Should i switch to OpenAI module or is there a solution to fix this?

Ive tried setting different models as the primary and fallback providers.

:clipboard: Error messages or input/output bundles

[
{
“result”: [
{
“error”: {
“message”: “Anthropic has returned an error: Invalid model claude-3-5-sonnet-latest. Available models are: claude-3-opus-latest, claude-3-7-sonnet-20250219, claude-3-5-haiku-latest”,
“type”: “ProviderException”
},
“status”: “fail”,
“provider”: “anthropic”,
“provider_status_code”: null,
“cost”: 0
},
{
“error”: {
“message”: “Provider has returned an Error. Please try again later”
},
“status”: “fail”,
“provider”: “openai”,
“provider_status_code”: null,
“is_fallback”: true,
“cost”: 0
}
]
}
]

:camera_with_flash: Screenshots (scenario flow, module settings, errors)

make-models-1.jpg
make-models-1.jpg1559×582 119 KB

3

Hi Riaz,

What you’re seeing is expected with the Eden module right now. Eden automatically selects the “latest” model for a provider, and in this case Anthropic deprecated claude-3-5-sonnet-latest, so Eden is still trying to call a model that’s no longer available. Since you can’t manually choose the exact model, the request fails before the fallback kicks in properly, which is why you’re seeing both providers fail.

There isn’t much you can fix on your side. Until Eden updates their internal model mapping, the safest option is to switch to the native OpenAI module, where you can explicitly select a supported model and avoid burning credits on failed runs.

Hopefully Eden updates this soon, but for now this is a provider-side issue rather than a configuration problem.

Hope this helps.

Regards, Tony