aura loss tweet. you cut them off, just let them save face and move on.

why in the world would anyone need to make that caveat? You guys took away access to your models, they said "we're sad but it was 5%", and now you need to say "Actually..." What's your point? You are hurting them more than they are saying?

This is a REALLY bad look for OpenAI. Grow up.

Maybe don't screw over ChatGPT customers who use Cursor then? No developer on earth wants you guys waging corporate warfare inside our computers.

This sucks. As good as it is for content this sucks for consumers and I hate this

Do you want to give them the model or not? If not, fine, but don’t gaslight them. Let bygones be bygones. Was it not enough that they sold your token for you? You’ve now rug pulled them, and you still expect them to explain the downside consequences of you firing them? It lacks class.

token efficiency may have some impact, but I think the major factor is that people are just using Fable and other models because they’re better

It's just a number at the end of the day. Regardless, who gets hurt in this situation is users. I think that's very clear and deserves more attention than any of this other stuff you all are throwing at each other. Users don't care about that.

Buddy we're either using xAI or Anthropic models. Don't feel bad

"Trust me what we did actually did hurt you" boohoo

Not to mention a huge chunk of us that love OAI models just go to you now directly :)

I’m a heavy Grok user but I still use and like Codex/ChatGPT. I love having one review the other’s plans and work. You see things I don’t see but … Your lead has evaporated Grok is at least competitive I don’t see how OpenAI and Anthropic will be able to maintain high prices at scale.

Tibo have you considered replacing cursor with Charleston AI?

Complete aura collapse You are not frontier, that belongs to Fable. And gpt models are actually replaceable Take the L until gpt 6 is ready bro

As a paid subscriber of both of you, I'm really disappointed by this change. I really enjoy OpenAI's models — I think GPT 5.6-Sol is the best daily use model right now and the second-best large planning model. But I think Cursor is a better harness. I hope y'all change this decision.

@thsottiaux Maybe people prefer using oai models in Codex or OSS tools via their oai subscription But cursor defaulting subagents and new users / auto users to composer prob also contributes to this

MY 5% IS NOT THAT SMALL, IT IS HOW YOU USE IT THAT MATTERS!! - says him while she leaves his house disappointed

What did you just try to imply that it’s the low number because ChatGPT is token efficient? Did I just actually see that with my eyes the model that in every single one of my tests takes by far the longest time and the most amount of tokens being said that it’s token efficient is why it’s only 5% used. I must be misunderstanding because what are you talking about? Go to artificial analysis and look at the hallucination rate for GPT models. It’s 98% GLM is 30 just stop please bro.

so basically you are going back to your ex

So size does matter?

You realize coming out in defense of something like this looks 100x more pathetic, right?

Bahaha .. Tibo, you hurt by the facts?? This is an epic cope post

Your company do shit then get called out a tiny part and that trigger you ? Man have a bit of respect and join cursor. You have been brainwashed

I don’t think you needed to respond to this lmao - rare Tibo L

I’m having a hard time understand why you would feel the need to say this

Don’t be pressed, just make codex the best way to code with openai models

Also: Codex >>>

You guys screwed up with this one. If you want to be a common utility, act like one.

Just kiss and make up already

Huh? You guys are the ones shutting them down and you want them to throw you a parade and flowers??? GTFO! 😂

Tibo I am very sorry but this I do not agree with, Cursor has been the main driver on AI assisted coding and OpenAI going as far as closing their models there because a little beef between two people is not good for humanity, I think the team should reconsider.

Tibo just let it go. OpenAI made their bed. Michael is justified in his response.

Do you not understand they’re a public company with shareholders to answer to? You created this mess for them despite them being your partner. You initiated it, not them. So is squeezing the last few ounces of narrative control on the situation really necessary? If this is the direction of travel, just get it done properly and move on. Anything else is pretty poor taste.