pricing plans | Voters

pricing plans

Phil Gibson

the pricing plans are a bit wonky imo. 
$10 should be something like 750 prompt and 1000 flow
$60 is just ridiculous. i used 500 within 1 week with somewhat heavy usage. i would never pay $60 out of spite as i outright refuse to obtain extra credits via flex because i feel that if im paying for a monthly subscription, then my credits should last a month. you need to be more realistic with your pricing. maybe something like 10, 15, 20, 25, etc so we can find the one that best suites us. and $10 for 300 prompts? bit of a joke mate.

December 17, 2024

erd online

Dear Windsurf Team,
I'm writing to share feedback about the Pro Ultimate tier's 3000 credit limitation. While I understand the need for usage tiers, I believe the current structure could be enhanced in one of two ways to better align with the "Pro Ultimate" positioning:
Remove Credit Limitations The "Ultimate" in Pro Ultimate suggests unrestricted access. Having a credit cap seems to contradict this premium positioning. True "ultimate" access would mean unlimited usage, allowing users to fully leverage the platform's capabilities without constantly monitoring credit consumption.
Enhanced Premium Features If credit limits must remain, consider enhancing the Pro Ultimate tier by:
Providing access to more advanced AI models
Offering priority compute resources for faster response times
Including exclusive features not available in other tiers
Providing higher quality code completions and suggestions
Adding specialized domain expertise or industry-specific optimizations
These improvements would make the "Ultimate" designation more meaningful and provide clear differentiation from other tiers, justifying both the premium positioning and any usage limitations.
The goal is to ensure users feel they're getting a truly premium experience that matches the "Pro Ultimate" name, whether through unlimited usage or through significantly enhanced capabilities.
Best regards

Judioo

For me the price is questionable because:
* Distribution of credits
* Pricing upgrade path (from $10 ->$60)
* Distribution of credits
You have Prompt & Flow Action credits. However as a user I have no way of controlling or knowing when and what my prompt consumes. I have observed that when I prompt and the result generates multiple actions I will consume flow credits. However, how much and when this happens I have no control over.
This becomes a problem, especially on the more recent models where (for some reason) they do not automatically start a flow actions, meaning I often am required to prompt again to initial a flow action. Consuming more credits than needed.
Whats worst is when either credit allocation expires and we are forced to purchase flex credits. The issue here is they cost the same weather you consume prompt or flow. This FEELS wrong as within my plan you give me 3 TIMES more flow action credits than prompt. Why then do you not  divide the flex credits in the same way?
It make everyone NOT want to purchase flex credits.
* Pricing upgrade path (from $10 ->$60)
The only reason for me to upgrade is if every month I consume over 3000 flex credits (i.e. $50 extra per month). However because of above it is not clear what the "advantage" of going up to ultimate is?
It simply isn't clear the advantage and "real" cost savings I would make upgrading to ultimate as the split credit system and the lack of control I have in choosing which credit I consumed makes upgrading a complicated decision.

Serhani Abdel

last month I had finished my month with 0 flow actions vs 187 prompt
which means that flow actions are consumed far than premium prompts 
Unlimited prompts on the Pro Plan it's just useless for me, if you can ask the premium models freely using the legacy mode?
we need flow actions, not premium prompt bro
Additionally you have suggested to make the $10 plan with 1000 flow actions & 750 prompt, honestly it's not sufficient

Friday Lee

I'd pay $100 a month for true unlimited

Duyvinh229148

the pricing is ridiculous, I used up in 2 week, I would spend $20 for cursor to last to end of month, or maybe create a new windsurf account to spend another $15. They think we're stupid with that $10 and $60 for 3000 flow actions

Maxime Troubat

totally agree. The rates make no sense. everything should be unlimited for $60 if you see the prices of competitors

翟贺

The thread starter is right. $15 is too little for 500 times, and $60 is a waste. If you don't want to increase the number of times $15, then you can provide a variety of price positioning options.

bspeagle

I'd pay $100 a month for true unlimited

Bhavesh

Phil just get better at prompting mate and use the base cascade model it works

Gevik Nalbandian

First. Great job with Windsurf. There are a few things to consider with pricing. One is that different users will have varying experience levels making them more or less prone to understanding errors generated by the models. I have found myself in both camps depending on my expertise on the programming language, env etc… what drains credits is the endless loops. So they is a bigger hole to cover. So in the meantime a more generous set of credits would be allowed. Secondly the jump from $10 to $60 is big and it would be useful to have a middle tier.

Bhavesh

Gevik Nalbandianhi Gevik I would say use the cascade models as they are really good. I'm halfway through the month and I've only used 300 cascades prompts for large stuff.

Gevik Nalbandian

Bhavesh I am not sure I fully agree. I do use the base model but it’s quite lacking. I find myself getting more out of my paid OpenAI ChatGPT than the base model. So along with pricing changes that I have suggested, being able to use other models (BYOapiK) would greatly help.

Bhavesh

Gevik Nalbandian Very strange, what are you building that it requires expensive model to just achieve the output?

Gevik Nalbandian

Bhavesh maybe we're talking about 2 different things. As we know Windsurf (paid version) allows for the use of Claude 3.5 Sonnet (known for it's coding capabilities). The Cascade base model, while it can answer questions and maybe make small changes, it does require a lot more "encouragement" - clearly there is a reason we have access to Claude 3.5 otherwise, why use it? 
Example: If you asked the base model to access your codebase and say add a new set of CRUD APIs that interact with a relational data model, it will do its best to tell you how "you" can achieve it, but it's a far cry from Claude 3.5 just jumping in and doing the work. 
Another example: when debugging a CSS issue, the base model went into circles (and believe me, I tried to give it the benefit of the doubt), but after numerous attempts, I fell back onto Claude 3.5 Sonnet, and in a few short attempts, we had the answer to the issue I was facing. 
As I said in my original post, the developer's experience matters. Claude 3.5 allows someone with limited experience to get far (and fast) without needing to know all the details of a language's syntax, available packages/libraries, etc. The whole point of Windsurf is to have an accelerated development environment.

Bhavesh

Gevik NalbandianI I see. it seems the people who uses short prompts relies heavily on claude and o4. I cannot help with that you might have to fork out $$$ to do those as you are offloading the work to the AI to figure out what you mean.

so far I've built a lot with base cascade model by providing proper prompts. The output is similar to o4/claude. so far I've had to only use claude 3.5 where there are recurring issues on logic based

→