Groq
Configuring Groq in Aimogen Pro for very fast open-weight model inference, which models are available, and the Groq Responses API caveat.
Groq runs open-weight models on custom hardware. Its selling point is latency: responses often arrive several times faster than from comparable hosted models.
Get a key#
- Sign in at console.groq.com.
- Create a key at console.groq.com/keys.
Groq has a free tier with per-minute limits, which is enough for evaluation.
Configure#
Aimogen Pro › Settings › API Keys › Groq AI API Keys
One key per line, Test API key, save.
Models#
Aimogen Pro 2.8.8 lists 16 Groq models.
| Model | Notes |
|---|---|
llama-3.3-70b-versatile | General-purpose workhorse |
llama-3.1-8b-instant | Fastest, cheapest, lowest quality |
openai/gpt-oss-120b, openai/gpt-oss-20b, openai/gpt-oss-safeguard-20b | Open-weight GPT models |
groq/compound, groq/compound-mini | Groq composite systems |
moonshotai/kimi-k2-instruct-0905 | Long-context |
qwen/qwen3-32b | Qwen 3 |
meta-llama/llama-4-scout-17b-16e-instruct | Llama 4 Scout |
meta-llama/llama-guard-4-12b, meta-llama/llama-prompt-guard-2-* | Safety and prompt-injection classifiers, not writers |
canopylabs/orpheus-* | Speech-oriented models |
Groq models are in the plugin function-calling capability list, so tool calling works. Groq is not in the vision list — image input is not supported.
When to use Groq#
Good for
- High-volume bulk generation where throughput matters
- Interactive chatbots, where latency is felt directly by the visitor
- Cheap first-pass drafting refined by a stronger model afterwards
Less good for
- Nuanced long-form writing. Open-weight models at this size trail the frontier closed models
- Vision tasks. Not supported
- Very long context. Windows are smaller than Claude or Gemini
The Groq Responses API#
Settings › Advanced AI Settings › Enable Usage Of Groq's Responses API (Experimental)
The plugin help text is explicit: Groq Responses API is in beta and does not yet support streaming, tool calls and other AI features. Leaving it off is the right default.
If you enable it and your chatbot stops streaming or extensions stop firing, that is why.
Common problems#
429 quickly on the free tier Groq free limits are per-minute and strict. Raise the request delay or upgrade.
Model returns a label instead of an article You selected a guard or classifier model. Choose a writing model.
Tool calls ignored Check whether the Groq Responses API toggle is on. Turn it off.
Endpoint used#
https://api.groq.com/openai/v1/chat/completionsRelated#
Still stuck? Open a support ticket and include the diagnostics from Aimogen Pro › System & Logs › System Info.