Way more excited about GPT 5.6 and Gemini 3.5 than Mythos

Latest rumors are Mythos may be as much as $400 per 1M tokens.

It's really a vanity model that isn't practical

GPT 5.6 and Gemini 3.5 Pro will be far more usable and early indicators are 5.6 is fantastic

Way more excited for GPT 5.6 and Gemini 3.5 Pro than Mythos

The rumored price of Mythos is now at $400 per 1M tokens!

Making it more of a vanity model than something that is usable

5.6 and 3.5 will be far more pragmatic, faster and useful

@bindureddy $400/1M tokens means Mythos is a flex for people who expense their API bills and never look at the invoice. Nobody shipping real shit is touching that. GPT-5.6 and Gemini 3.5 are the actual workhorses here.

@bindureddy TRUE

@bindureddy @grok How much you charge for $1M tokens ?

@bindureddy But what if the Mythos output justifies the $400 per 1M cost

@bindureddy bindu, received five of these a/b tests today while using GPT-5.5 Pro. the 'new version' alternate responses are fast and smart. hope this is an indication that 5.6 is on the way

@bindureddy Mainly excited for gpt 5.6 at this point, since can use 3rd party harnesses with it, it really feels worth it as main driver.

As for mythos, if prices are as high expected then effectively useless for me, but maybe they release distilled version

@bindureddy $400/M makes it a research model, not an infra model. Honestly if 5.6 delivers on the early signals, it probably captures the agentic use case anyway.

@bindureddy I'm still more excited about Gemini 3.5 Pro

@bindureddy Wow That's really good Would have capped in but no coins yet

@bindureddy Right now all this Mythos drama looks like marketing hype. Every AI lab will have a superior model internally but they are projecting it like the God tier. $400/1M keep your model Dario, No thanks!

@bindureddy "Agreed. If those pricing rumors are true, Mythos looks more like a showcase model than something most developers can use at scale. GPT 5.6 and Gemini 3.5 Pro seem far more likely to hit the sweet spot of performance, cost, and real-world adoption." 🚀

@bindureddy Mythos at $400/M is peak 'we built it but forgot who it's for' energy.

@bindureddy Whats scary is the open source models that will come out after they reverse engineer how mythos works

@bindureddy agree

@bindureddy Have you tested GPT 5.6 already?

@bindureddy $400 per million tokens is a research model pretending to be a product honestly

@bindureddy $400/1M tokens for a benchmark flex? Models that win are the ones teams can actually ship in production.

@bindureddy Vanity model is the wrong frame. o3 got the same "too expensive" treatment and serious use cases emerged. Wait for actual evals before calling it.