Groq shut down Llama 4 Scout and Qwen3 32B on its free and developer plans. Their model IDs were meta-llama/llama-4-scout-17b-16e-instruct and qwen/qwen3-32b, and Groq announced the shutdown on June 17. On the free plan Llama 4 Scout had allowed 30,000 tokens a minute and Qwen3 32B 60 requests a minute, each with 500,000 tokens a day. Groq named gpt-oss-120b as the replacement for Qwen3 32B, and gpt-oss-120b or Qwen3.6 27B for Llama 4 Scout.
Source ↗