TwiScan
Hot
Communities
Account collections
Login
Register
English
日本語
한국의
简体中文
繁体中文
Register and share your invite link to earn from video plays and referrals.
Register now
LinearUncle
@LinearUncle
服务器技术专家,曾就职于阿里云 ,有赞,外企,在线教育等; 👑 AI coding - 职业工程师 💻 只分享硬核AI编程工具,技巧等
Joined March 2020
1.4K
Following
14.3K
Followers
LinearUncle
@LinearUncle
2026.09.07 10:27
OpenAI 这刀法,真够鸡贼的 😂 这个小哥测了一下:同样是 GPT-6 Astra,订阅里的输出速度,比直接调 API 慢了很多。 普通模式: • 订阅:约 36 tokens/s • API:约 81 tokens/s,超过订阅的 2 倍 快速模式: • 订阅:约 71 tokens/s • API:约 124 tokens/s 开了快速模式的订阅,还跑不过普通模式的 API。 反观 Anthropic,很良心啊,Fable 5.1 在这次测试中,订阅和 API 都在 94 tokens/s 左右。
Show more
Youssof Al Toukhi
@Youssofal_
2026.09.07 07:17
I tested the speed of Astra and Fable 5.1 on the API and plan. - Astra is over 50% slower compared to the API. (81 tps vs 36 tps) - Fable speeds are identical. Another example of how sneaky OpenAI is.
Show more
0
0
45
43
0
Forward to community
Most Popular Users
Tibo
@thsottiaux
745.8K Followers
Serenity
@aleabitoreddit
1M Followers
Donald J. Trump
@realDonaldTrump
111.9M Followers
Elon Musk
@elonmusk
241.7M Followers
Sam Altman
@sama
6.3M Followers
Wall St Engine
@wallstengine
195.8K Followers
zerohedge
@zerohedge
3.4M Followers
OpenAI
@OpenAI
5.4M Followers
Polymarket
@Polymarket
2M Followers
Kalshi
@Kalshi
470.1K Followers
Bitcoin Archive
@BitcoinArchive
1.8M Followers
SpaceXAI
@SpaceXAI
2.1M Followers
Gavin Baker
@GavinSBaker
348.8K Followers
Grok
@grok
9.1M Followers
Anthropic
@AnthropicAI
1.8M Followers