Deepseek v4.1 flash uses a lot of the Other models usage

Is this an error? It’s a very cheap model but somehow it fills up the Other models usage really fast. I was using the high variant and used it up in few days. I’m on the old Pro plan, maybe it just gives a little of the usage for other models?

3 posts were merged into an existing topic: Any plans to add DeepSeek V4.1 Flash to the Cursor Token Rate exclusion list?