Groq 是什么？
Groq builds its own inference chips (LPUs) and offers GroqCloud, an API serving open models at very high token speeds with low latency.
That speed matters for voice assistants, real-time agents and anything where users wait on the response.
主要功能
- Very high tokens per second
- Low latency
- Open models
- OpenAI-compatible API
- Free tier
