OpenAI to Launch GPT-4.5 Model in the Coming Weeks, Confirms Altman

393
13 Feb 2025
6 min read

News Synopsis

OpenAI is on the verge of releasing its next artificial intelligence model, GPT-4.5, following previous development setbacks. CEO Sam Altman confirmed on social media platform X that the model, internally referred to as "Orion," will be available within weeks. Reports from Bloomberg in late 2023 suggested that Orion initially failed to meet OpenAI’s expected performance benchmarks. However, the company appears to have made significant progress and is now ready to introduce it to the public.

GPT-4.5: The Last Model Without Advanced Reasoning Capabilities

The launch of GPT-4.5 will mark a turning point for OpenAI. Altman stated that this will be the final model that does not incorporate additional computing power to simulate human-like reasoning. The company has already started integrating this advanced reasoning approach into newer AI models, such as o1 and o3.

With AI technology evolving rapidly, OpenAI aims to transition towards models that can process information in a more human-like manner. These improvements are expected to enhance AI’s decision-making capabilities, making responses more context-aware and precise.

A Unified Approach: Merging GPT Models with O-Series AI

Altman announced that OpenAI will soon unify its GPT models with the o-series AI to streamline user experience. Traditionally, OpenAI users have had to select from multiple model versions, each offering different capabilities. This complexity has often led to confusion among users about which model best suits their needs.

By integrating the GPT models with the o-series, OpenAI aims to develop an AI system capable of determining how much computational power it should use for a given task. This means that instead of users manually selecting a model, the AI itself will assess and allocate resources efficiently, ensuring faster and more relevant responses.

Making AI More User-Friendly

Acknowledging the complexities of current AI model selection, Altman emphasized OpenAI’s commitment to making AI more intuitive. He admitted that the "model picker" feature—where users choose between different versions of GPT—has become cumbersome. OpenAI's goal is to simplify the process so that AI interactions feel seamless, requiring minimal input from users.

“We want AI to ‘just work’ for you; we realize how complicated our model and product offerings have gotten,” Altman wrote in his post. This move aligns with OpenAI’s broader vision of making artificial intelligence more accessible to businesses, researchers, and everyday users.

Looking Ahead: OpenAI’s Plans for GPT-5

Beyond GPT-4.5, OpenAI has set its sights on the next major AI evolution—GPT-5. Altman revealed that OpenAI is actively developing the new model, which is expected to integrate many of the company’s existing technologies, including o3.

A smaller version, known as o3-mini, has already been introduced. However, OpenAI has yet to unveil the full-scale GPT-5 model. If successful, GPT-5 could mark a significant leap in AI’s ability to reason, adapt, and provide highly accurate responses.

Final Thoughts: OpenAI’s Future in AI Development

With the imminent release of GPT-4.5, OpenAI is closing one chapter and opening another in AI advancement. The company’s shift toward human-like reasoning in AI and its efforts to simplify user interactions signal a transformative phase in artificial intelligence development.

As OpenAI moves forward with GPT-5 and its broader AI initiatives, the world will be watching closely to see how these advancements reshape the AI landscape.

Podcast

TWN Special