Groq Component : error to fetch models LLM

[Anthony-Perriolat]

Bug Description

When I add component Groq and i refresh list model to field model, i had errors strange and i havn't list to models llm on groq.
I have only 2 models llama 3.1 8B and 70b

Errors when i fetched :

Error while updating the Component

    Error code: 400 - {'error': {'message': 'The model canopylabs/orpheus-arabic-saudi does not support chat completions', 'type': 'invalid_request_error'}}

Error while updating the Component

    Error code: 400 - {'error': {'message': 'tool calling is not supported with this model', 'type': 'invalid_request_error', 'param': 'tool calling'}}

Error while updating the Component

    Error code: 400 - {'error': {'message': 'The model canopylabs/orpheus-arabic-saudi requires terms acceptance. Please have the org admin accept the terms at https://console.groq.com/playground?model=canopylabs%2Forpheus-arabic-saudi', 'type': 'invalid_request_error', 'code': 'model_terms_required'}}

Error while updating the Component

    Error code: 400 - {'error': {'message': "Failed to call a function. Please adjust your prompt. See 'failed_generation' for more details.", 'type': 'invalid_request_error', 'code': 'tool_use_failed', 'failed_generation': '<function=test_tool>{"x": "test"'}}

I tried to change url base by "https://api.groq.com/openai/v1/" but noting happen

Error while updating the Component

    Error code: 400 - {'error': {'message': "Failed to call a function. Please adjust your prompt. See 'failed_generation' for more details.", 'type': 'invalid_request_error', 'code': 'tool_use_failed', 'failed_generation': '<function=test_tool>{"x": "test_data'}}

Log docker :

2025-12-21T08:53:25.796292Z [info     ] Found 21 models from Groq API 

2025-12-21T08:53:25.796490Z [info     ] Testing 13 LLM models for tool calling support...

2025-12-21T08:53:25.944033Z [info     ] Client error during session scope: 400: Error code: 400 - {'error': {'message': 'tool calling' is not supported with this model', 'type': 'invalid_request_error', 'param': 'tool calling'}}

Reproduction

Install with docker -> docker run -p 7860:7860 langflowai/langflow:latest
make poject and add component Groq.

Expected behavior

I would like get list models with gpt OSS 20b

Who can help?

@ogabrielluiz - @nicoloboschi - @zzzming - @jordanrfrazier - @mfortman11 - @NadirJ

Operating System

Ubuntu linux 24.04

Langflow Version

latest => 1.7.1

Python Version

3.12

Screenshot

No response

Flow File

No response

[coderabbitai]

📝 CodeRabbit Plan Mode

Generate an implementation plan and prompts that you can use with your favorite coding agent.

• Create Plan

Examples

• Example 1
• Example 2

---

🔗 Similar Issues

Related Issues

• GROQ API for LLM not working , getting component building error due to Proxies #4983
• Error Building Component - Docker #5470

👤 Suggested Assignees

• giridharg96
• Gabrieln18

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord or schedule a call!

[MugheesMehdi07]

Problem Summary

When refreshing the Groq model list, users were seeing errors and only 2 models (llama 3.1 8B and 70b) instead of the full list. The errors shown were:

• "The model canopylabs/orpheus-arabic-saudi does not support chat completions"
• "tool calling is not supported with this model"
• "The model ... requires terms acceptance"
• "Failed to call a function"

Root Cause

The issue was in the _test_tool_calling method in groq_model_discovery.py. The code was only catching a limited set of exceptions:
n
except (ImportError, AttributeError, TypeError, ValueError, RuntimeError, KeyError) as e:However, the Groq Python library raises groq.APIError (and potentially other API-specific exceptions) when:

• A model doesn't support chat completions
• A model requires terms acceptance
• Tool calling fails
• Other API-level errors occur

These APIError exceptions weren't being caught, causing them to propagate up and either:

1. Crash the model discovery process, or
2. Get logged as unhandled errors, making it appear that the discovery failed

The Fix

I've made two key improvements:

1. Enhanced Exception Handling in _test_tool_calling

• Changed: Now catches Exception (which includes all Groq API exceptions)
• Added: Specific error message detection for the exact error types mentioned in this issue:
  • Models that don't support chat completions
  • Models that don't support tool calling
  • Models requiring terms acceptance
  • Tool use failures
• Improved: Changed log level from warning to debug for expected errors (reduces log noise)

2. Added Defensive Error Handling in Model Discovery Loop

• Wrapped: Each model test in its own try-except block
• Benefit: If one model fails catastrophically (e.g., unexpected error in _get_provider_name), it doesn't stop testing other models
• Result: All models are tested and included in the final list, even if some fail

---

Could I please be assigned to this issue? The fix is ready and tested. I'll create the PR once assigned.

[andifilhohub]

@Anthony-Perriolat

Thank you for your report!

[andifilhohub]

LFOSS-3144

[Anthony-Perriolat]

thanks lot for your help. @andifilhohub @MugheesMehdi07
It's a nice project open source, i will give a some time for help this project. Tag me if you need help for anything.
Thank you again.

[andifilhohub]

Fix: #12158

Thanks for the report, @Anthony-Perriolat .