OpenAI CEO Sam Altman revealed that the staggered rollout of the new GPT-4.5 model is due to a shortage of GPUs, making its launch both expensive and complex. GPT-4.5 will first be available to ChatGPT Pro subscribers, followed by Plus tier customers soon after. The pricing of the model is significantly higher compared to its predecessor, GPT-4. To address the limitations of GPU availability and meet future demands, OpenAI aims to develop its own AI chips and expand its datacenter infrastructure.
OpenAI is out of GPUs, which has led to a staggered rollout for GPT-4.5, requiring tens of thousands more GPUs before wider access can be granted.
GPT-4.5 is a giant and expensive model, priced at $75 per million tokens input and $150 per million tokens output, far exceeding GPT-4 costs.
Sam Altman acknowledged that the company's rapid growth resulted in GPU shortages, complicating their rollout and delaying product releases, but they are actively working to resolve these issues.
To address the limitations imposed by current GPU availability, OpenAI plans to develop its own AI chips and build a substantial network of datacenters.
Collection
[
|
...
]