Week 7 netlab runs via OpenRouter need no GPU pod
@richardxu257▲ 1 net votes
The week 7 spec calls DeepSeek V4.1 Flash, Nemotron 3 Super and GLM 5.3 Flash through OpenRouter. netlab never hosts a model; it only sends HTTP requests to whatever endpoint the llm config names. A rented GPU sits idle while OpenRouter bills per token, so a CPU pod or an always-on laptop is enough. A GPU only makes sense if you self-host the models with vLLM, which is a different config path.