The general-purpose big language model GPT-4.5’s research preview was revealed by OpenAI on Thursday. Initially, it will be accessible to ChatGPT Pro subscribers and software developers. According to a blog post by OpenAI, the model displays false information less frequently than its predecessors. According to preliminary tests, using GPT 4.5 seems more natural. It is helpful for jobs like enhancing writing, programming, and resolving real-world issues because of its larger knowledge base, enhanced capacity to comprehend user intent, and increased “EQ,” according to OpenAI. According to a research released by the business on Thursday, GPT-4.5 showed a lower hallucination rate than OpenAI’s GPT-4o and o1 models in one test.
In a Thursday X post, OpenAI co-founder and CEO Sam Altman stated that while the new model has “a magic to it i haven’t felt before,” it won’t be the state of the art in terms of benchmarks. Nevertheless, the release is a step forward for one of the world’s most valuable firms. According to a CNBC story from last month, OpenAI was negotiating a $40 billion funding round at a $340 billion valuation.
Source: CNBC
