骑墙看远

[骑墙看远 @ysli_cn]

@songkeys 一整天了，不敢用astra，我是plus，sol都不够用

[骑墙看远 @ysli_cn]

@thsottiaux with so biggest breakthrough, 2 bank resets just not enough

[骑墙看远 @ysli_cn]

@Raphael1402X @grok done, why are you doing this?

[骑墙看远 @ysli_cn]

@xueyu1125 @ZixuanLi_ glm模型本来就是国产最强，可惜zcode app拉的不行

[骑墙看远 @ysli_cn]

@thsottiaux one more thing, how fast will astra eat the credis compare to sol?

[骑墙看远 @ysli_cn]

@thsottiaux thank you, but i was expecting more bank resets

[骑墙看远 @ysli_cn]

@xueyu1125 muse就是专门训练的考题，解决问题和sol根本不在一个档次

[骑墙看远 @ysli_cn]

@uncleTomat0 这是ai吧

[骑墙看远 @ysli_cn]

@mylifcc glm不是一直慢的出名吗

[骑墙看远 @ysli_cn]

@thdxr hi, could you pls help understand something that omen alpha is $100 usage out of $10, but we're only given 57,900 prompts per month, much less than muse spark 1.3 contributor which is %60 usage out of $10

[骑墙看远 @ysli_cn]

@thsottiaux reaaaally hate the 5h limitation, it turns 2h work into whole day long

[骑墙看远 @ysli_cn]

@shuai1721 @ZaiforStartups 中国用户不配

[骑墙看远 @ysli_cn]

@zhibo0502 @Fei2411 glm舔美舔习惯了

[骑墙看远 @ysli_cn]

@thsottiaux just take your days pls

[骑墙看远 @ysli_cn]

@ZixuanLi_ 🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢🐢

[骑墙看远 @ysli_cn]

@zstmfhy 不过有没有可能openai在全球搞几个独立的计算中心，这样就算一个挂了，还有别的可以用

[骑墙看远 @ysli_cn]

@zcode_ai 没有你说的这个卡片，重启也没有

[骑墙看远 @ysli_cn]

@thsottiaux @ChatGPT gpt disconnected, we need a reset, do we?

[骑墙看远 @ysli_cn]

@Vincent_AINotes 前两周可以，现在已经没了

[骑墙看远 @ysli_cn]

@_kaichen 输出是挺快，1.3比1.2变强了很多，但是用下来两个问题：1.间歇性只说不做；2.长文本中文语义理解、推理能力比deepseek v4 flash感觉略差，但是差距可以接受