You're all wrong, $2000 ChatGPT Max is coming
And you will like it
Speakers are now announced for Latent Space LIVE! at NeurIPS.
The announcement of the $200/mo ChatGPT Pro plan yesterday was met with an incredible amount of skepticism. Variously: $200/mo is a decoy. It’s a con. Open source will catch up. $200/mo is crazy given $20/mo competition. Sonnet is better1. Even sama felt compelled to clarify that Pro is just for “a small percentage of users” and for “the hardest problems”, while the blogpost stresses its “4/4 reliability” scoped to “researchers, engineers, and individuals who use research-grade intelligence daily”.
Though we have made light jabs at loss projections in the past2, the pricing power of OpenAI is a very pertinent matter to the entire industry whether or not you are a fan. If OpenAI cannot successfully launch a $200/mo prosumer product that stays meaningfully ahead of the competition, it may struggle to justify further research into more powerful models like Orion, and fail to finance the road to AGI.
Because online sentiment is so one-sided, we’d like to make the case for the opposite: not only do plans for $2000/month per seat ChatGPT subscriptions make sense, they’ll probably happen as soon as next year. It’s the probable outcome if you take AGI seriously and actually believe that “2025 is when agents will work” and don’t just spin your wheels counting letters in strawberries.
Back in September we were first to put prices on each level of OpenAI’s AGI roadmap interpolating from $20 to the rumored $2000 and then 2 OoMs beyond:
Keep reading with a 7-day free trial
Subscribe to Latent Space to keep reading this post and get 7 days of free access to the full post archives.