> ## Content Index
> Fetch the complete content index at: https://www.testingcatalog.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# OpenAI expands o3 capabilities with higher limits and task scheduling
- URL: https://www.testingcatalog.com/openai-expands-o3-capabilities-with-higher-limits-and-task-scheduling/
- Published: 2025-04-23T13:12:21.000Z
- Updated: 2025-04-23T19:13:10.000Z
- Author: Alexey Shabanov
- Tags: ChatGPT News, Latest AI News

Update: o3 rate limit confirmed to be doubled to 100 messages per week. 

Reports from multiple users indicate that OpenAI may have lifted the o3 model’s quota from 50 messages per week to about 50-100 messages per day. The company has not confirmed this shift, so it is still possible the recent uptick reflects a one-time quota reset rather than a new permanent ceiling.

> Looks like ChatGPT o3 and o4-mini limits got increased! Is it 50 a day now? [https://t.co/lJNQb9XNWv](https://t.co/lJNQb9XNWv?ref=testingcatalog.com) [pic.twitter.com/qaiub5lEUR](https://t.co/qaiub5lEUR?ref=testingcatalog.com)
> 
> — TestingCatalog News 🗞 (@testingcatalog) [April 23, 2025](https://twitter.com/testingcatalog/status/1914978759653921110?ref%5Fsrc=twsrc%5Etfw&ref=testingcatalog.com)

The same chatter hints at an upcoming “o3 Pro” tier that would outclass the older o1 Pro in handling complex workloads. If so, the higher limit could be an early sign of that paid upgrade, giving testers a taste of the capacity boost before any formal launch.

> Confirmed 🔥  
>  
> o3 and o4-mini rate limits have been doubled! [https://t.co/dAGE2ut597](https://t.co/dAGE2ut597?ref=testingcatalog.com) [pic.twitter.com/uVYhDEmhZN](https://t.co/uVYhDEmhZN?ref=testingcatalog.com)
> 
> — TestingCatalog News 🗞 (@testingcatalog) [April 23, 2025](https://twitter.com/testingcatalog/status/1915110213822148884?ref%5Fsrc=twsrc%5Etfw&ref=testingcatalog.com)

Meanwhile, both o3 and the lighter o4-mini now support the task-scheduling feature that was once limited to a special 4o test model. OpenAI is reportedly preparing incentives to drive adoption, and the standalone “4o with tasks” option is expected to disappear as scheduling becomes a standard capability across models.

> New ChatGPT web app announcement: "Use o3 and o4-mini to schedule tasks"  
>  
> However, the tasks scheduled using o3/o4-mini still seem to use the default 4o model for executing the scheduled tasks in the background [https://t.co/4CeJDjle6U](https://t.co/4CeJDjle6U?ref=testingcatalog.com) [pic.twitter.com/ValLpYZeGC](https://t.co/ValLpYZeGC?ref=testingcatalog.com)
> 
> — Tibor Blaho (@btibor91) [April 17, 2025](https://twitter.com/btibor91/status/1912756414402928723?ref%5Fsrc=twsrc%5Etfw&ref=testingcatalog.com)

[ChatGPT](https://www.testingcatalog.com/tag/chatgpt/) remains the main doorway through which users will experience these backend changes.