BrianChen

@letsvpn 真是操你们妈了

【关于账户时长冻结、自助退款及后续计费安排的说明】 我们正在把账户冻结、退款状态查询和自助退款流程逐步做成可操作、可查询的系统功能，尽量减少等待和反复沟通。关于关键时间线、免费服务时长补发及 5 月 8 日后的计费安排，我们已整理为完整说明，请点击查看详情： letsvpn.world/blog/dba00f

@letsvpn ？这什么鸡巴毛？我被影响了一个月现在说我不在回补范围内？？

@letsvpn @Yoyoxgx 什么时候退啊，怎么不说话装死

@Yoyoxgx 可以成功退款的，请不必担心这个问题

4 月 27 日 最近很多人都在议论"是不是跑路了"。理解大家的担心，换作我们自己也会问。但事实是：团队全员在岗，在不断寻找+测试可用方案。问题会解决的，请再给我们一点时间。

@AOXVPN 能不能打游戏

大家好！ 给大家介绍下小牛VPN！ 1.团队全员base在新加坡，请放心使用！ 2. 不保存用户数据，免邮箱注册，完全匿名！ 3.付款是通过第三方支付方式，购买计划不用担心有暴露风险！ 4.希望小牛VPN能够让大家更加自由的上网！ 小牛VPN团队致

@mark_k @arena @OpenAI more like frontend leaderboard

GPT-5.5 @arena ranking: Solid but not dominant placement for @OpenAI GPT-5.5 High on Code Arena. It's sitting comfortably in the top 10 (#9 with an Arena score of 1,500), which is respectable for a new model release. However, it's noticeably trailing the current coding leaders, mostly Anthropic's Claude Opus 4.x family.

@cholf5 @caizhenghai 5.5low就和5.4high的思考程度差不多了，还贵，5.5high肯定更贵跑的更快。

@caizhenghai 对，5.5 high

Codex 这个额度是认真的吗，跑一个逆向任务直接5h清空，weekly limit 从 100% 用到 71%，要是这么玩的话，我可要粉转黑了。

所以我的 #ChatGPT 是个小姐姐

@BropeJim @SIGKITTEN What’s even weirder is that the reasoning depth on Low also feels comparable to 5.4 High. Not sure if that’s a bug or intended behavior

@SIGKITTEN weird some random people were saying 5.5 works better on LOW reasoning and it overthinks too much on higher? i dunno havent had time to test much

is it just me or is 5.5 medium basically regarded

@BropeJim @SIGKITTEN I managed to fully reverse-engineer an app in under two hours using Low, and it cost me just over 20 from my weekly quota. Honestly, 5.5 Low feels almost as expensive as 5.4 High in terms of usage. .

@frxiaobei 用low一个小时跑完5小时额度...

GPT-5.5 这个价格有点美丽。 不过， OpenAI 给出的解释是 GPT-5.5 的 token 效率更高， 完成同样的 Codex 任务需要更少的 token， 所以实际成本未必翻倍。

👀

@letsvpn 用不了了已经

4 月 23 日 我们不会放弃 全团队都在盯着，都在想办法。请大家放心，不管多么困难，我们都不会丢下这一切不管。

@JasonBotterill 👀

OpenAI has updated codex model list so gpt-5.4 no longer says “latest frontier”

@TheAmolAvasare ok boomer

@LexnLin Thursday

What did i tell you guys GPT 5.5 tomorrow ;)

@kyliebytes tongiht i have had a couple of drinks x.com/sama/status/20…

@Suolar_ @scaling01 🫡

@chetaslua @gdb 5.4thinking

🚨 OpenAI PRO Vision test according to @gdb gpt 5.5 is newly trained model and its initial checkpoint for SPUD > highly effecient > they poured everything they learnt in last 2 years > there is no wall Prompt : find the stocking in the image and mark it , more tests coming

@nicdunz ewwww

fleshy fruits by image 2

@chetaslua @trendcrusher7 omg

@Cheney_NehcOal @trendcrusher7 this one was 5

XBOX 360 ASCII ART Sparks of AGI > spud cooking hard