OpenAI has doubled the rate limits for GPT-4o and GPT-4-mini-high models for ChatGPT Plus subscribers, allowing users to send and receive more messages per hour. This update, announced by CEO Sam Altman on X, is designed to improve the overall user experience by providing more flexibility for those who rely on these models for high-frequency tasks.
Feedback led to the update
The announcement comes after feedback from users who had raised concerns about the limitations on usage. Altman acknowledged that OpenAI is making an effort to listen to this feedback in order to improve the service.
Despite this update, OpenAI highlighted that the company's ongoing work involves balancing user demand with technical limitations.
GPU shortages
OpenAI also pointed out that the company is still facing challenges with its infrastructure, especially when it comes to the availability of GPUs, which are required to run these AI models. Altman explained that OpenAI has to make “hard tradeoffs” between things like rate limits, adding new features, and managing latency.
The demand for these advanced AI models is high, and managing system performance at scale remains a challenge for the company.
In addition to the recent increase in rate limits, OpenAI will remove its GPT-4 model from ChatGPT on April 30 and fully transition to GPT-4o, which is now the default model for ChatGPT users. Altman noted GPU shortages as a key challenge in the phased rollout. To address this, OpenAI planned to add tens of thousands of GPUs to meet demand.
Feedback led to the update
The announcement comes after feedback from users who had raised concerns about the limitations on usage. Altman acknowledged that OpenAI is making an effort to listen to this feedback in order to improve the service.
Despite this update, OpenAI highlighted that the company's ongoing work involves balancing user demand with technical limitations.
GPU shortages
OpenAI also pointed out that the company is still facing challenges with its infrastructure, especially when it comes to the availability of GPUs, which are required to run these AI models. Altman explained that OpenAI has to make “hard tradeoffs” between things like rate limits, adding new features, and managing latency.
The demand for these advanced AI models is high, and managing system performance at scale remains a challenge for the company.
In addition to the recent increase in rate limits, OpenAI will remove its GPT-4 model from ChatGPT on April 30 and fully transition to GPT-4o, which is now the default model for ChatGPT users. Altman noted GPU shortages as a key challenge in the phased rollout. To address this, OpenAI planned to add tens of thousands of GPUs to meet demand.
