GPT-6 Astra on $100 Codex Pro plan: faster and cheaper on quota
A user reports Astra Light feels much faster and Astra Max ran 41 minutes using only 6% of weekly quota, versus simple tasks eating 15% a week earlier on the $100 Codex Pro plan.
Wait, is GPT-6 Astra finally getting faster nd more usage efficient??? It used to be a worst model nd now On the $100 Codex Pro plan it's complete opposite. Just a week ago, a few simple Astra tasks could consume 15% of weekly quota in minutes. Now? - Astra Light feels incredibly fast, even without 1.5x speed. - Astra Max ran for 41 mins and consumed only 6% of weekly usage. That's a pretty massive difference. Could OpenAI's recent caching and efficiency improvements finally be paying off or they've improved Astra??? Anyone else noticing Astra becoming faster and cheaper lately, or is this just an isolated experience??