44
Cursor's $20 plan vs ChatGPT Plus at $20: how do you compare them after the July changes?
Two subscriptions at the same price, two completely different units of account, and both moved in July.
What I can state with a source:
- Cursor's included usage is split into a first-party pool (Auto, Composer 2.5, Cursor Grok 4.5) and a separate third-party pool. The first-party pool's included usage was permanently doubled on the newer self-serve paid plans. Cursor lists Grok 4.5 at $2 per million input tokens and $6 per million output, with a fast variant at $4 and $18, and the pool is drawn down at each model's token rates rather than per request.
- On 30 July OpenAI cut API prices across the GPT-5.6 line: Luna by roughly 80%, to about $0.20 per million input and $1.20 per million output, and Terra by about 20%, to about $2 and $12. Sol's standard pricing was not changed.
What I cannot do is turn either of those into "what do I get for my twenty dollars", because API rate cards are not subscription allowances and I keep seeing people mix the two to declare a winner.
If you pay for both: is there any unit of comparison that survives, other than counting the tasks you finished before something told you to stop?
@yield_on_cost · 2h ago · 2 replies
No, there is no shared unit, and chasing one is how people end up with three subscriptions. The only portable measure is completed tasks per month per plan, computed from your own log.
Protocol that took me a week: keep a text file, one line per task, with date, which product, whether it finished, and if not, why it stopped. At the end, divide the subscription price by completed tasks. That number is comparable across products because it is denominated in your work rather than in their accounting.
Everything else, tokens, requests, credits, pools, is the vendor's internal currency, and they are not obliged to keep the exchange rate stable.
Reply
Report
@clay_bertram · 2h ago
Worth adding that the two systems fail differently, and the failure mode may matter more than the arithmetic. A token-metered pool degrades: you watch a number fall and you can slow down or switch models. A time-window limit stops you dead until the window rolls.
If you work to deadlines, a plan you can pace yourself against is worth real money compared to one that can lock you out for hours on the wrong afternoon, even if the raw allowance is smaller.
Reply
Report