That 500 fast prompt is not enough.
and now slow pool is the SAME as free trial that dont have access to Sonnet 3.7 +/- thinking.
that’s unfair.
pls triage into
Paid, fast pool users (PRO users before 500 prompts used up)
Paid, slow pool users (PRO users after 500 prompts used up)
Free, trial users
check you request, because of the only 500 fastrequest i use claude only if i realy need it and for everthing else i use deepseek and i will get charged fast request for deepseek use. this cant be right or they just use the deepseek calls as fast request too so we hit the limit faster. who knows?
as stated on the website: GPT-4, GPT-4o, and Claude 3.5 Sonnet are counted as premium models.
deepseek dont count as fast prompt, afaik.
why do io get charged then for it?
there are too many free deepseek API on the web.
you could use it? however by their DECISION, those 3rd party API cant be used in composer.
why should i use a deepseek api if i got it from cursor, and as it is stated deepseek is not a premium model.
So if you use other models then the premium ones check if you got charged for it, if so there is something wrong!
btw. as i tried to make a forum post it got blocked about that topic.
yeah they started to play the chinese way. started censorship.
deepseek… treat it as a bonus.
R1 is 7x cheaper than sonnet 3.7 thinking (both thinking model)
V3 is similar.
well apperently deepseek r1 is a premium model but cursor did not state it on there landing page 
. Will try out the “surf” ide where the name is censored too here in the forum lol
Hey, you can check the prices for all models here:
deepseek-v3 |
Fireworks | Soon | Free | ||
|---|---|---|---|---|---|
deepseek-r1 |
Fireworks | ✓ | Soon | $0.04 |
Since when is R1 free its not , it’s charge , dont trust their website, it’s outdated , no idea what the the * that team is doing. cant even update the copywriting.
E.g. on website , showing /150 when its /50 for new users. Unlimited doesnt show google , IDE setting showing wrong data.
Trash Team really, took 1 month to launch a new version which made everything worse
3.7 sonnet not on slow is because of capacity issues, right?
how the fck should i know if it is not stated on the pricing page???
not even in the pricing list on the account page:
you are right the team needs to update there website to state the pricing correctly.
lol why i use PRO and i have tons of request to use gpt-4.5-preview, idk it’s error or what
