OpenAI is paving the way for its next major AI evolution, confirming the development of GPT-5, a model the company says is designed to unify and streamline its increasingly complex lineup of AI offerings.
This strategic shift aims to create a more powerful, multimodal, and accessible AI experience for all users. As a ‘stepping stone,’ the company introduced GPT-4.5 as a research preview on February 27, 2025, internally codenamed "Orion."
ChatGPT-4.5 “Orion”
ChatGPT-4.5 was trained with unprecedented compute power and data for improved pattern recognition and creative insight.
While offering better performance and efficiency than GPT-4, OpenAI notes it's not considered a "frontier" model and serves as the final release focused on traditional processing without the advanced, chain-of-thought style reasoning planned for its successor.
ChatGPT-5: Everything users need in one model
GPT-5 marks a departure from maintaining separate models like GPT-4, GPT-4 Turbo, and specialized o-series models (including o3).
Instead, it will merge the best capabilities into a single, comprehensive system designed to intelligently utilize all of OpenAI's tools.
In other words, the model will inherently decide when to engage in deep reasoning, provide quick responses, or leverage integrated tools such as search, code interpreter, memory, voice interaction, and the new Canvas feature for visual reasoning.
The goal, as OpenAI states, is to move away from user complexity and deliver AI that "just works."
This approach promises a more intuitive and seamless user experience, most notably by eliminating the often-confusing model picker. GPT-5 aims to embody "magic unified intelligence," instinctively knowing the appropriate level of processing or which tool to employ for a given task.
Delivering AI that 'works'
we are getting things under control, but you should expect new releases from openai to be delayed, stuff to break, and for service to sometimes be slow as we deal with capacity challenges.April 1, 2025
Recently, OpenAI introduced users to image generation directly within ChatGPT-4o, which brought both excitement and frustration. The model develops exceptionally realistic images, but at incredibly slow processing times, especially within the free tier.
OpenAI first rolled out the feature to all users but scaled it back to only Plus and Pro users due to overwhelming demand. However, despite technical difficulties that even included ChatGPT outages, the company has made it free once again.
It’s seems as if OpenAI has their work cut out for them with managing how to get ChatGPT-4o image generation to work before rolling out the new model. Perhaps that’s why CEO Sam Altman has not given a specific launch date for ChatGPT-5 other than “soon.”
Accessible to all
Accessibility is a key part of the GPT-5 rollout plan. All ChatGPT users, including those on the free tier, will have access to GPT-5 at a "standard intelligence" level.
Paid subscribers (Plus and Pro tiers) will benefit from higher intelligence settings capable of handling more complex requests and utilizing advanced features like deep research capabilities, dynamic voice interactions and real-time visual reasoning.
No release date yet
While a specific release date for GPT-5 remains unconfirmed, OpenAI's roadmap suggests a launch in the coming months.
The introduction of in-chat image generation, GPT-4.5 and the development of the unified GPT-5 signal are significant milestones, promising users worldwide more powerful, integrated and user-friendly AI tools ahead.
More from Tom's Guide
- I let ChatGPT Tasks schedule my day — and it was a disaster
- I just found out what happens when you feed Gemini 2.5 ChatGPT’s best prompts — it completely added its own spin
- I told ChatGPT something that I still regret — here's 7 things you should never share
![HIDevolution [2024] ASUS ROG...](https://images.fie.futurecdn.net/products/848664f20a82da37ee7b66f813eb40cdef2cfcf5-100-80.jpg.webp "HIDevolution [2024] ASUS ROG...")
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
