redocly 了解一下，自行管理

自家的 junie 还行

@catfly 次数差不多，但是新收费策略，经常给限速，用的不爽。

Pro: ~225 Sonnet 4 requests, ~550 Gemini requests, or ~650 GPT 4.1 requests

Pro+: ~675 Sonnet 4 requests, ~1,650 Gemini requests, or ~1,950 GPT 4.1 requests

Ultra: ~4,500 Sonnet 4 requests, ~11,000 Gemini requests, or ~13,000 GPT 4.1 requests

@Chatterleys 四楼回答了

@1016 元宝帮我写的邮件，你自己改下。

Subject:​​ Request to Revert to Old Pricing – Annual Subscription Holder

Dear Cursor Team,

I am writing as an annual subscriber under my previous pricing plan. As confirmed in your recent communications:

"The pricing you purchased will continue to be reflected for the duration of your billing period."

​However, I was unable to opt out of recent changes​ and wish to revert to the ​request-based pricing model​ until its sunset, as indicated in your policy:

"Existing users wanting to use request-based pricing [...] can reach out to our team ([email protected])."

Please assist me in transitioning back to this legacy plan for the remainder of my annual billing cycle ([End Date, if known]). I appreciate your support in honoring this option for existing users.

​To simplify your process:​​
🔹 ​Account Email:​​ [Your Account Email]
🔹 ​Subscription Tier:​​ [Pro/Basic, if applicable]

Thank you for your prompt attention to this request.

Best regards,
[Your Full Name]
[Contact Information, if needed]

@xubingok 有错过的，比如我，现在页面上已经没有回退按钮了。

@NNNNzs 应该是的，我的回退生效时间，也是到下次付费前。

@creating2000 选国区付费后，还能切换到其他区吗，us 区的 ai 可以使用 claude 模型

我在大陆地区，用的招行的 visa 。+ 1

云电脑，手机也有 app ，凑合玩玩呗。

java 性能一般？啥叫性能好

https://github.com/danielgtaylor/huma 好奇，为啥没推荐这个的呢，看着不错，不过我也没用过。

@ylh1024 还没换 60 刀，先去测测 jb 的 junie

@steptodream 直接砍到一半不到了，https://docs.cursor.com/account/pricing#expected-usage-within-limits

Expected usage within limits

Expected usage within limits for the median user per month:

Pro: ~225 Sonnet 4 requests, ~550 Gemini requests, or ~650 GPT 4.1 requests
Pro+: ~675 Sonnet 4 requests, ~1,650 Gemini requests, or ~1,950 GPT 4.1 requests
Ultra: ~4,500 Sonnet 4 requests, ~11,000 Gemini requests, or ~13,000 GPT 4.1 requests
Models like Opus 4 consume more tokens per request and reach usage limits faster than other models. We recommend selectively and intentionally choosing these models.

@Brocol 对，其他模型写的，和免费的，比如千问、ds 之类的都差不多了，太恶心了

@JYii 可以手动指定模型，不喜欢 auto ，手动指定的 claude 4

@shiguangweiming 我年费的也没这个选项了，找不到了。

@phithon 感谢，我去试一下