OpenAI has launched GPT-4.5, its latest AI model, initially available as a research preview for ChatGPT Pro users. Described as OpenAI’s “most knowledgeable model yet,” GPT-4.5 offers better writing, improved world knowledge and a more refined personality. It also enhances computational efficiency over GPT-4 by 10x and hallucinates less than previous models.
Despite these improvements, GPT-4.5 is not a frontier model and performs below o1 and o3-mini in some evaluations. OpenAI trained it using new supervision techniques, SFT and RLHF. CEO Sam Altman called it a “giant, expensive model” that won’t break benchmarks.
GPT-4.5 will expand to Plus and Team users next week, followed by Enterprise and Edu users. OpenAI is also planning to launch GPT-5 as early as May, incorporating its o3 reasoning model for more advanced capabilities.
Source: The Verge
