OpenAI blog hints new model is similar size; caching and inference gains double intelligence per dollar

eliebakouch · x · 2026-09-23

Citing OpenAI's blog line that "improvements in caching and inference let us serve these models at lower cost," the author infers the new model is likely similar in size to its predecessor, with gains coming from inference-side optimization. Either way, it's a big win for users: intelligence per dollar has more than doubled under the new pricing.

Original post →

More from Models

Models channel →