However, from a moral and ethical point of view, I can not support my money being passed to down to (a) specific company/ies cursor offers models from.
So my question is: how is the income from Pro subscriptions split and spread across the companies behind the available models?
E.g., can I be assured that, when only using model 1 from company X, company Y does not see any cent from my subscription fees?
I can understand you’re most likely not willing to disclose your contracts with the model’s companies, but maybe some bits of info, as in: cursor pays a flat rate or per queries/tokens/other-usage or whatever measures.
The Pro plan includes $20 of API usage per month. When you use a specific model (say, Claude from Anthropic or GPT from OpenAI), you’re charged at that model’s published API token rates - and those costs are deducted from your included usage budget. Cursor charges at published API rates with no markup. So your usage-based costs are directly tied to which models you actually use.
If you use Auto mode, Cursor selects models automatically behind the scenes - and you’re charged at fixed rates ($1.25/1M input tokens, $6.00/1M output tokens) regardless of which model is selected. In that case, you don’t control which provider handles your request.
The practical takeaway: if you pick a specific model from a specific provider, your token usage goes toward that provider’s API costs. If you want to avoid funding a particular company, stick to models from other providers and don’t use Auto (since you can’t control what it picks).