Ask HN: Is ChatGPT-o1 underlying model better or just using more inference time?
Is chatgpt-o1 just using more time during inference and passing it through more routines than the underlying model being better at reasoning?
But beyond that they're keeping secret.