Groq publishes a limit for each model, and the binding one for chat is usually tokens, not requests. At 8,000 tokens per minute, a single request with a 6,000-token prompt uses most of the minute; at 200,000 tokens per day, a chatbot averaging 2,000 tokens per exchange stops after about a hundred exchanges, long before the 1,000 requests per day.
Only one model in the free table accepts images (qwen/qwen3.8-27b), and each image is billed as 2,048 input tokens: three images in one request already take 6,144 of the 8,000 tokens per minute.