Having been “lucky” enough to get my Claude account banned again today, I really feel the need to vent a little about this company.
Why do I say “lucky”?
Because they refunded my money, which effectively means I got to use it for free for over half a month. That’s worth at least $150.
My account, which had been stable for over three months, has officially bitten the dust.
Although I could easily just open a new account and subscribe again, I turned around and spent $200 on GPT instead.
This company is genuinely hilarious. The founder came from Baidu; I wonder how Baidu tortured him to make him so anti-H (anti-Chinese/anti-Asian).
Their initial bet was on the coding sector—an area where Large Language Models excel and where commercialization is easiest to implement. The model’s capabilities go without saying; it has been going back and forth with OpenAI. That was until their API usage was eventually surpassed by ChatGPT.
Subjectively evaluating the Opus 4.6 model, its programming ability no longer holds a clear advantage over the previous version of GPT Codex 5.3. Furthermore, it is even weaker in instruction following. Its only remaining advantage (sounding more human) completely vanished after the release of GPT 5.4.
I hope they release Opus 5.0 soon so I can return to Claude.
By the way, next time I might just buy an annual subscription. They can ban me all they want, as long as they keep issuing refunds.
This company, constantly fighting against itself, is truly absurd. I have this unrealistic fantasy: I really want to work in the risk control department at Anthropic. I wonder how their performance is measured. Do they get more bonuses the more accounts they ban and the more money they have to refund? Where else could you find such a sweet gig?