Over the next week, we'll starting rolling out two small improvements to our pricing based on your feedback:

1. You'll be able to pay for usage-based models with your requests.
2. MAX, our advanced option for enabling very long context models, will shift from per-tool-call pricing to API pricing.

#1 will simplify our pricing and lower costs for our users. Instead of having to turn on usage-based pricing to purchase models like o3, you'll be able to use the requests you already have included in your subscription.

#2 will improve the transparency and fairness of our pricing. Charging per tool call meant that low context requests were very over-priced (and long context ones were under-priced). API pricing allows cost to flex up and down naturally with the amount of context used and aligns incentives.

Please let us know if you have any questions or feedback! And thank you for being vocal. These changes were highly requested, and we're excited to make them available.