OpenAI aims to teach its “open” AI model to ask for help from cloud models

According to TechCrunch, the new functionality will allow an open model to directly access the OpenAI API to access the company's other, larger models, which will significantly increase its computing power.

According to the source, the idea for the transfer function was proposed by a developer during one of OpenAI’s recent developer forums. The suggestion appears to have gained support within the company. OpenAI is holding a series of developer feedback events to help shape a future release of the open model.

The on-premises model, which can leverage more powerful cloud systems, resembles Apple Intelligence, Apple's suite of AI capabilities that uses a combination of on-device models and models running in "private" data centers.

According to TechCrunch’s sources, OpenAI is in the process of developing a new model from scratch, rather than repurposing an old one for public release. The open model is expected to be worse than the o3 model, but outperform the R1 DeepSeek counterpart in certain metrics.

The company recently launched its most advanced models, the o3 and o4-mini, which have significantly better performance in math and reasoning tests compared to previous OpenAI AI models.

Read the country's main IT news in our Telegram

On the topic

Read the country's main IT news in our Telegram

OpenAI is interested in buying Chrome if Google is forced to sell its web browser

On the topic

OpenAI is interested in buying Chrome if Google is forced to sell its web browser

OpenAI requires developers to provide ID to access their AI models to avoid copying by DeepSeek

On the topic

OpenAI requires developers to provide ID to access their AI models to avoid copying by DeepSeek

OpenAI launches budget-friendly Flex Processing API mode that halves costs

On the topic

OpenAI launches budget-friendly Flex Processing API mode that halves costs