Speculations are buzzing in the tech community as a potential leak hints at OpenAI’s latest iteration, GPT-4.5. Shared on various social media platforms, the leak, if accurate, unveils impressive features and pricing structures that could revolutionize the landscape of large language models.

GPT-4.5 Overview

GPT-4.5, the purported upgrade to OpenAI’s renowned GPT LLM, seems to introduce multi-modal capabilities across vision, video, audio, language, and 3D. The leak, initiated by user daniel_nyugenx on Twitter and discussed in a Reddit thread, emphasizes the model’s potential for complex reasoning and cross-modal understanding. However, skepticism looms as the authenticity of these claims remains unconfirmed.

Pricing Breakdown

According to the leaked draft, GPT-4.5 comes with a new pricing structure that raises eyebrows. The model is speculated to cost $0.06 per 1K tokens for input and $0.18 per 1K output tokens. A detailed breakdown includes variants like GPT-4.5 64K and GPT-4.5 Audio and Speech, each with its pricing tier. Notably, these prices surpass the existing GPT-4 rates, leading to discussions about potential implications for users and developers.

Community Response and Skepticism

As news of the leak circulates, the tech community is divided in its response. Some see this as a groundbreaking moment, anticipating a paradigm shift in content creation possibilities. However, others raise questions about the credibility of the leak, considering the ease with which misinformation can be fabricated online. Comments on the original Reddit thread reflect the uncertainty surrounding the accuracy of the pricing and draft.

OpenAI’s Response and Future Prospects

OpenAI CEO Sam Altman later confirmed on X that the “leak” was not real. Whether OpenAI is going to release GPT-4.5 or jump straight to GPT-5 is unknown. The next model would be the successor to GPT-4, which was released on March 14, 2023.

After GPT-3 was released in June 2020, it took two years to release GPT-3.5, which came out in March 2022. Meanwhile, OpenAI is already working on GPT-5. In July, the AI firm filed a trademark application for GPT5, which includes software for AI-based human speech and text, converting audio into text and voice and speech recognition.

In November, OpenAI CEO Sam Altman told the Financial Times the company is working to develop GPT-5, though he did not commit to a timeline for its release.

Our Say

In the wake of the alleged GPT-4.5 leak, the tech community was abuzz with anticipation, contemplating the potential advancements. However, OpenAI CEO Sam Altman promptly debunked the leak, emphasizing its inaccuracy. This revelation casts doubt on the speculated features and pricing, urging a cautious approach. While the possibility of GPT-4.5 remains uncertain, Altman’s confirmation of the ongoing development of GPT-5 adds a layer of intrigue to OpenAI’s plans. As enthusiasts eagerly await official updates, it underscores the importance of relying on verified information in navigating the evolving landscape of advanced language models.