Photo credit: www.cnbc.com
OpenAI has unveiled the research preview of its latest model, GPT-4.5, making it available initially to software developers and ChatGPT Pro subscribers. This announcement was made during an event in Davos, Switzerland, where CEO Sam Altman emphasized the advancements that this new iteration brings to the realm of artificial intelligence.
According to OpenAI, the new model presents inaccuracies less frequently than its predecessors, indicating a significant improvement in reliability. In a blog post, the company stated that early user interactions with GPT-4.5 feel more intuitive. The model’s enhanced knowledge base, improved capacity to comprehend user intent, and heightened emotional intelligence (EQ) contribute to its effectiveness in various tasks, including writing assistance, programming, and problem-solving. OpenAI also noted that GPT-4.5 showed a reduced rate of generating incorrect information compared to earlier models as detailed in an accompanying report.
Despite these enhancements, Altman acknowledged that GPT-4.5 does not represent the pinnacle of benchmarking in AI, describing it as possessing a unique quality that he had not experienced previously. This release marks another step forward for OpenAI, recognized as one of the most valuable startups globally. Recently, CNBC reported that the company is negotiating a substantial funding round valued at approximately $40 billion, which would solidify its market position further.
Read more CNBC reporting on AI
OpenAI plans to extend access to GPT-4.5 to ChatGPT Plus and Team subscribers shortly, with expectations to launch next week, as confirmed by Alex Paino, the firm’s research lead, during a livestream. Following this, subscribers to ChatGPT Edu and Enterprise will be granted access the week after.
Altman noted that GPT-4.5 is a resource-intensive model, indicating challenges in meeting user demand: “We really wanted to launch it to Plus and Pro at the same time, but we’ve been growing a lot and are out of GPUs. We will add tens of thousands of GPUs next week and roll it out to the Plus tier then,” he explained.
Additionally, GPT-4.5 is available through Microsoft’s Azure AI Foundry, as announced by CEO Satya Nadella on X. Microsoft, which has invested over $13 billion in OpenAI, plays a crucial role by providing computational resources and integrating OpenAI’s models into its suite of products.
This launch follows a period of market volatility spurred by advancements from competitors like DeepSeek, which led to a notable decrease in the market capitalization of Nvidia, a key supplier of chips for AI model developers. Just weeks later, Altman remarked on the company’s intent to enhance transparency regarding its development roadmaps. He indicated that after GPT-4.5, a future GPT-5 model would leverage a broader array of OpenAI’s technologies, while also noting that GPT-4.5 would be the final model classified as a “non-chain-of-thought” model.
WATCH: Distillation and the future of the Softbank, OpenAI partnership
Source
www.cnbc.com