Groq

Groq (note: not xAI's Grok) is a LPU hardware acceleration, ultra-low latency specialized inference service that mainly hosts open-source models such as Llama / Mixtral / Whisper, and its output speed is usually several times that of ordinary cloud inference.

Get an API Key

• Go to GroqCloud Register account
• API Keys → Create API Key, copy gsk_... Key

Configure in Cherry Studio

• Open Settings → Model Services, find Groq Provider to enter the details page
• API key Fill in gsk_...
• API address Default https://api.groq.com/openai/v1, no need to modify
• Click Get model list

Recommended usage

Model	Suitable scenarios
llama-3.3-70b-versatile	General chat, extremely fast
llama-3.1-8b-instant	Simple tasks, millisecond-level response
mixtral-8x7b-32768	Long context
whisper-large-v3	Speech to text

Applicable scenarios

• Real-time chatbot: Groq's "near-instant response" is very suitable for IM integration (paired with channel)
• high concurrency: tokens per second are significantly higher than ordinary cloud inference
• Don't care about having the latest model: Groq mainly hosts open-source models such as Llama, and does not have closed-source models like GPT-5 / Claude-4

Distinguishing Grok vs Groq

	Grok	Groq
Company	xAI (Musk)	Groq Inc.
Focus	Self-developed large models + web access	LPU hardware + open-source model inference
Model	grok-4 and other in-house	llama-3.x and other open-source

Grok (xAI) and Groq are often confused. In the Cherry Studio Provider list, they are two separate entries, so please pay attention to the distinction.

---

💡 Get help and submit feedback

If you encounter any questions, bugs, or have suggestions for feature improvements during configuration or use, please refer to Feedback and Suggestions the official channels provided there.