正在加载视频...

视频加载失败

🚨 codex users!! this is what you are getting for $20 GPT-6.1 Sol at Max (using fast) > 37 Million tokens > 53 Minutes of work > 48% of your WEEKLY usage!!! > $18 API usage Its EXTREMELY SLOW on normal The worst part is you are only getting...

57,436 次观看 • 7 天前 •via X (Twitter)

48 条评论

Edi Vilarinho 的头像
Edi Vilarinho7 天前

Estressei o GPT-6.1 Sol Máximo ao limite no projeto Umbra hoje e os números do monitoramento são surreais. 🚀 Em apenas 191 requisições, o modelo processou um total de 27.9 MILHÕES de tokens. O grande destaque? 96% de taxa de acerto de cache no input (27.7M). Otimização absurda de contexto. 🧵👇 Mas o ritmo cobrou o preço: a intensidade de geração de código foi tão alta que torrei a cota de 5 horas em apenas 3h e 2min (ritmo de 33%/h). Foram 193.6K tokens de saída, com 49.5K dedicados só ao reasoning. O modelo voa, mas é preciso monitorar de perto para não ficar sem cota no meio do flow! 😅💻

shownotover 的头像
shownotover7 天前

Yeah man, you got such a large amount of tokens only because the model is so slow. Look at how many tokens it processed in under 3 hours. Opus 5.5 would have processed 60 million tokens in 1 hour. It's not good being a Codex user now.

Edi Vilarinho 的头像
Edi Vilarinho7 天前

Outro caso que eu testei 30/09 - 14:50 - sol 6.1 high 30/09 - 19:54 - sol 6.1 high 1.5x 01/10 - 00:54 - sol 6.1 high

Sahil Panhotra | Indie Builder 的头像
Sahil Panhotra | Indie Builder7 天前

so jusy 36$ weekly in api usage 🫠

shownotover 的头像
shownotover7 天前

kinda... First it was $360 / month... then $300 ... and now it has come to this

Sahil Panhotra | Indie Builder 的头像
Sahil Panhotra | Indie Builder7 天前

🫠

shownotover 的头像
shownotover7 天前

I used for FASTEST and CHEAPEST scarping service... they support MCP server so... Best for my research tbh

VulKan 的头像
VulKan7 天前

that is a rough usage and what?!! 12tps..🫠

shownotover 的头像
shownotover7 天前

yeah... its really really really slow

Shahzeb Naveed 的头像
Shahzeb Naveed7 天前

codex is nerfed...

shownotover 的头像
shownotover7 天前

yeah

orvian 的头像
orvian7 天前

switching to claude asap

shownotover 的头像
shownotover7 天前

I bought $100 max 1.5 days ago... absolutly amazing!

James Malsawm 的头像
James Malsawm7 天前

Thank for the warning in advance. For me $20 on Claude is the best.

shownotover 的头像
shownotover7 天前

Yeah man, I bought $100 Claude max too

Zekun 的头像
Zekun7 天前

Bro is still using GPT models? Why not embrace Opus/Fable 😭 I tested the tok/s on Sol and Astra, and they’re painfully slow compared to Opus/Fable.

shownotover 的头像
shownotover7 天前

I bought Claude Max $100 just 1.5 days ago bro

Zekun 的头像
Zekun7 天前

Perfect timing 😭

LEO 的头像
LEO6 天前

As much as I hate OpenAI/Anthropic, anyone who uses MAX or even HIGH to make a point is a dumbass. You're not CERN or NASA evaluating something deep. You're not even the Supreme Court or a tiny biotech lab. You're just a doritos eating hobo creating a meme. Use slow or stfu.

Kekius Pisimus 的头像
Kekius Pisimus7 天前

Delete Codex

shownotover 的头像
shownotover7 天前

I can't... I paid $20

Kekius Pisimus 的头像
Kekius Pisimus7 天前

I know you're a big fan of Scam Altman 😂

Frank T. Mackie 的头像
Frank T. Mackie7 天前

If you want real usage than maybe you should upgrade to a real Codex plan and stop bitching

shownotover 的头像
shownotover6 天前

Oh my... What a beautiful statement... Through my tests (all figures public) Claude $20 > Codex $100 right now

Peter 的头像
Peter6 天前

So using fast does not really worth it, right?

shownotover 的头像
shownotover6 天前

no... not at all

Artem 的头像
Artem6 天前

I've never use gpt-6.1 on max 😅

shownotover 的头像
shownotover7 天前

This is crazy bro... I wanna cry too

CuriousDev 的头像
CuriousDev7 天前

I guess they can't afford the compute cost now and offering cheap inference is not adding up

shownotover 的头像
shownotover7 天前

They are making bad after bad decisions... I loved codex in the past... now... I don't know why they are doing all this

CuriousDev 的头像
CuriousDev7 天前

Bro tbh codex earlier was also best because of the repeated resets, which they cant anyway sustain

vineeth 的头像
vineeth6 天前

Hey bro, Increase the work you can get done with those tokens by employing tools and skills like RTK, ponytail, caveman etc. This harness incorporates all of them and more- it increases the work you can get done with those 300 Million tokens by like 2x

shownotover 的头像
shownotover6 天前

thanks, good point, I know cavemen, should test them all! 🙏🏻

vineeth 的头像
vineeth6 天前

Glad to know i helped

NVortex 的头像
NVortex7 天前

Do you recommend I stick with the €20 Codex plan or switch to Claude at €20/month?

shownotover 的头像
shownotover6 天前

Switch to Claude... wayyyy better (I have tested both, if you don't trust my word, check my profile)

NVortex 的头像
NVortex6 天前

Thanks for your response! Are the limits the same, though? I’m asking because I rarely hit the limit with Codex.

Noah 的头像
Noah7 天前

48% of the weekly limit in 53 minutes is brutal 😭

shownotover 的头像
shownotover6 天前

it gets worse bro... trust me

زويد 的头像
زويد6 天前

It’s a good thing that I’m using Hermes for easy shifting. Bye Bye Codex Welcome Back Claude

Anon🇨🇭 的头像
Anon🇨🇭6 天前

engagement farming classic

CODIFY 的头像
CODIFY6 天前

Thanks for sharing this!

King of Cups 的头像
King of Cups6 天前

My expectations has been completely different.

𝓖𝓪𝓶𝓮 𝓴𝓝𝓲𝓰𝓱𝓽 𝓟𝓵𝓪𝔂𝓼 的头像
𝓖𝓪𝓶𝓮 𝓴𝓝𝓲𝓰𝓱𝓽 𝓟𝓵𝓪𝔂𝓼6 天前

Using Luna on high is really good! Never hitting my 5hr limit anymore!

yocoo 的头像
yocoo6 天前

快速模式,周额度 48% 的量,居然有 3700 万token。那么普通模式周额度可以有 1.9 亿token。看起来还可以的样子

shownotover 的头像
shownotover6 天前

Check Claude Pro $20 That gives you 2 BIllion tokens per month... 500 Million per week for Opus 5.5 (Top model in the world)

yocoo 的头像
yocoo6 天前

clsude的额度确实慷慨

P3xill 的头像
P3xill6 天前

How do you remove the 5h limit?

相关视频

I tested Claude Code on a fresh account - 1,500 lines of HTML cost me 50% of my window. Full video and summary is here.. I just ran a recorded test on Claude Code with a fresh account (Pro, not Max - my main account was 20x Max) , and the result is honestly insane. The task was trivial: create 3 simple demo HTML pages, around 500 lines each. Roughly 1,500 lines of code total. Nothing massive. Nothing enterprise-grade. Nothing that should meaningfully stress a premium coding product. And yet Claude Code burned through 40% of my 5-hour window almost immediately. I ran the exact same test with Codex, and it consumed only 2%. Then it got even worse: after the session ended, I did absolutely nothing for 15 minutes, and Claude still ate another 10%. Total: 50% of the 5-hour window gone for a tiny HTML demo. My weekly usage had already started at 2% before I even really used it, and after this tiny test it jumped to 8%. Now let us be generous and assume this entire run used around 30k tokens total. If 30k tokens represents 10% of weekly usage, that implies around 300k tokens per week. That is roughly 1.2M-1.3M tokens per month, and even if you round up aggressively, you are still in the 1.5M token range. Using the Sonnet 4.6 pricing you list: $3 per 1M input tokens $15 per 1M output tokens How exactly is this supposed to make sense for a paid coding product? Because from the user side, this no longer looks like "premium usage protection." It looks like a quota system that is either wildly inefficient, badly broken, or being accounted in a way users are not being told about. And that is before I even get to my main account: my $200 Max plan now dies in a single day. Just a few months ago, similar or heavier usage would last me about a week. So no, I do not buy the "maybe you just used it more" excuse anymore. Something is clearly broken in Claude Code. Either token accounting is broken, context handling is broken, background consumption is broken, or all three. Alex Albert is this really the experience you want users to pay for? Just watch the video. I tried to be very transparent and clear for your team! I was fan of Claude but just disappointed! And if you want, send me the detailed token accounting for this session and let us inspect it together publicly. Because from where I am standing, this is no longer a small pricing annoyance. It looks like something seriously wrong is happening, and users deserve a real explanation.

Hayrettin Tüzel

26,854 次观看 • 6 个月前