I'm about to start working on a token co-op for kimi 3 / qwen3.8-max when it's released. Setup aibrix in vultr and start selling coding plan seats.
mr. TIMseems like there’s huge incentive to be a model inference provider right now the prices listed by the lab are basically the going rate, so any optimizations you can make are pure additional profit OpenAI just showed that you can point Sol at an inference stack and save a ton of money