Since the creator is inactive, I've brought forward the close date just so that this market isn't forgotten about (traders - including me - will see it in their closed markets after it closes).

The date will be extended if OpenAI are yet to release a more capable model.

@chrisjbillington
OpenAI has recently released o1 and now announced o3, but they've confirmed in a blog post recently too that they plan to continue the GPT series of models and that the o-series is simply separate.

But wouldn't this technically resolve with both? Because it is technically true that a model with a new name is now better than GPT-4. But it seems like it will also be true that a GPT-4.5 or 5 model will release as well.

So this seems to actually fulfill both of these criteria for "yes" and "no"
-More capable successor is released with a different naming scheme.
- GPT-4.5 or GPT-5 (or similar) is released.

If a GPT-4.5 comes out next year, then o1 wouldn't be considered the successor right? or perhaps this market should already be resolved since OpenAI confirmed that the GPT series will be continuing separate from o-series.

@LDJ I've bumped the close date further to decrease the chances people think this is definitely going to resolve based on the state of things at the end of 2024.

As for the question, I have a position here so shouldn't be the arbiter of what the criteria mean. But I think it's pretty fair that if a GPT-X is released and it looks like a continuous chain of models in a series (as opposed to, say, GPT-6 being released and openAI saying "well, o1 was basically GPT-5" or whatnot), then a NO resolution would make sense.

This part of the criteria:

A bit of ambiguity exists with the word "successor" in the sense that a sufficiently different agentic model might be released, but it has to be their most "general" model and beat GPT-4 in most respects other than cost or speed.

Seems to explicitly anticipate this kind of scenario, and so it sounds like a sufficiently general replacement for GPT-4 would count. I'm open to people arguing that o1 already satisfies this, but would lean against. They're not marketing it as a general model and haven't given it internet access, image-generation capabilities, advanced voice mode, or advanced data analysis. Granted many of these things didn't exist for GPT-4 when this market was created, but it's a point at least that it's not their general replacement for GPT-4.