Lisan al Gaib, creator of the LisanBench benchmark, considers subscribing to ChatGPT Pro due to restrictive Claude rate limits

ahhhhhh

i love you anthropic i promise i will be back when you guys release mythos or have better rate limits

(I will only use the Pro sub for a month. Just want to finish a few projects)

@scaling01 openai loses money on pro sub so naturally i have one and it’s amazing

@AndrewCurran_ i will miss claudia for writing

@scaling01 Breaking Up Is Hard To Do.

@scaling01 Qwen has OUTSTANDING limits, literally almost NO limits. By the way, Qwen 3.7 Max is smarter than GPT 5.4 xhigh.

@scaling01 This image is AI-generated. Here's proof.

@_SandyBay_ good to know that SynthID doesn't work lmao

@scaling01 From my experience ($20 plan on both) limits are better with Claude.

@Presidentlin @iamgroguu Anthropic is still spiritually #1 with Mythos

but it's not here and they didn't release Claude 5, sooo

@scaling01 Two days before the end of 2x limits promo, lmao.

@scaling01 yeah the limits on Claude are brutal 😤

Pro is worth it if you're hitting walls constantly Lisan, that o1 access hits different

@scaling01 20x is amazing, but sometimes not enough lol

@R2Cdev_ $100

200 seems overkill for me, its a big jump from 20 to 200

@iamgroguu @scaling01 Happy but only temp, victory is when Lisan has to keep extending or renewing their sub.

It must be demonstrably clear that Anthropic is 2nd or 3rd for most tasks other than more likes on Twitter vs OAI (they kill them here)

@scaling01 I've used Opus A LOT for both coding and brainstorming. 4.8 is something else. It's almost A-sychophant. No matter how I try to shape my argument for something I'd like to be true, it just goes on pushing back, no matter what. And you go gpt? Now?!?! I get the ratelimit argument

@scaling01 Or you just generated your ChatGPT subscription with Nano Banana 2. I think my variant is much more likely.

@scaling01 @AndrewCurran_ Guys. I don't know about coding (yet) and I don't care about ratelimits. Opus 4.8's vending bench scores are justified. It will just... hold his stance. It reminds me of Opus 3, in some weird way.

Please just make sure that you aren't subconsciously NOT wanting that.

@scaling01 @Presidentlin the evil has been defeated!