GPT-4.5 Expected Next Week, with GPT-5 Looming Ahead
OpenAI’s Upcoming AI Models: A Look Ahead
OpenAI is gearing up to launch an updated version of its AI model, known as GPT-4.5, which is expected to debut shortly. Alongside this, the company is working towards what they hope will be their most significant advancement yet—GPT-5. There are high expectations for these releases, particularly regarding their potential to bring us closer to achieving Artificial General Intelligence (AGI).
What to Expect from GPT-4.5
Reports from various sources, including The Verge, indicate that OpenAI could release GPT-4.5 as early as next week. The anticipation builds as Microsoft has plans to host this new model, although an official announcement may take a bit longer. Also, speculation suggests that GPT-5 could make its appearance around May, coinciding with Microsoft’s developer conference, Microsoft Build, scheduled for May 22.
OpenAI’s Promises and Features
Sam Altman, the CEO of OpenAI, recently stated on social media that ChatGPT users will enjoy unlimited access to the "standard intelligence setting" of GPT-5 at no cost once it launches. One of the more notable features integrated into GPT-5 is the o3 reasoning model, designed to enhance the system’s fact-checking abilities. This improvement is particularly significant given OpenAI’s aim for broad reach and user accessibility with its upcoming releases.
Microsoft’s Role in the Launch
Microsoft has an incentive to launch GPT-5 in May, particularly due to previous advancements showcased during last year’s Microsoft Build event. In 2022, they introduced the Copilot+ line of PCs, setting a standard for on-device AI performing tasks. Moreover, Microsoft continues to enhance its Copilot experience, embedding AI features across various apps within Microsoft 365, GitHub, Dropbox, and even Windows 11.
Access Levels for Current Subscribers
Altman mentioned that current Plus subscribers will access a "higher level of intelligence" with GPT-5, while Pro subscribers are set to experience an even more advanced version. Meanwhile, OpenAI has been developing other products like Deep Research, which focuses on generating long-form reports based on initial prompts and follow-up questions. Although these features are expected to roll out more widely, users might encounter AI that tends to lean on resources like Wikipedia rather than thoroughly verified information.
Transitioning to GPT-4.5: The Last of its Kind
GPT-4.5, also referred to as Orion, is notable for being the last non-chain-of-thought model from OpenAI. This term refers to the ability of AI to decompose complex problems into manageable parts, analogous to human reasoning methods. The integration of the o3 reasoning model with the latest version could allow OpenAI to make claims related to milestones in AGI, although it’s essential to note that OpenAI’s definition of AGI may differ from broader public perceptions of true intelligence.
Skepticism Surrounding Revolutionary Claims
While some may view this next version as groundbreaking, it’s wise to be cautious regarding the actual advancements. Although GPT-5 may achieve improved reasoning capabilities, the critical concern remains whether these innovations will lead to new, practical applications of AI that fundamentally enhance how users interact with the technology.
Competitive Landscape
The competitive atmosphere has recently intensified following the unveiling of DeepSeek, a Chinese AI model developed at a fraction of the cost of GPT-4, yet demonstrating comparable or superior performance. OpenAI and Microsoft have criticized DeepSeek, claiming that it lacks originality, which brings a layer of irony given that GPT itself is trained on vast amounts of copyrighted content without permission. As the anticipation builds for the release of GPT-4.5 and GPT-5, OpenAI faces pressure not only from users but also from investors, eager to see impactful results that confirm the advancements outlined.
