OpenAI’s GPT-4.5 launched as a research preview, granting limited access to Pro users at $200/month and Plus users at $20/month. While CEO Sam Altman describes it as a more natural conversational model, it lacks major reasoning advancements.
Despite improvements in accuracy, hallucination reduction, and persuasion, GPT-4.5’s high cost—$75 per million input tokens and $150 per million output—has sparked debate. AI experts remain divided on whether its enhancements justify the price.
Andrej Karpathy compared GPT-4 and GPT-4.5 in an experiment where users judged five creative writing tasks. Surprisingly, GPT-4 won four out of five. Similarly, Dr. Raj Dandeker’s technical tests found little advantage, as GPT-4.5 struggled with math and logic, contradicting OpenAI’s claims.
Media reactions reflect mixed sentiments. Wired criticized OpenAI’s push for AGI, calling GPT-4.5 expensive with marginal gains. Futurism noted declining hype, while Stanford’s Jacob Rintamaki praised its humor as a leap in AI’s social understanding.
Even rival AI models chimed in. xAI’s Grok highlighted GPT-4.5’s conversational improvements but called it resource-intensive. ChatGPT itself emphasized better context retention, creativity, and accuracy but admitted flaws in long conversations.
Overall, GPT-4.5 refines AI’s linguistic ability but remains costly. While it enhances engagement and natural dialogue, its lack of substantial reasoning improvements leaves many questioning its value. As a research preview, future updates may define its place in AI’s evolution.
