OpenAI CEO Sam Altman has provided a major update on the company’s roadmap for its next-generation AI models, revealing that GPT-4.5, codenamed “Orion,” will be the final model before OpenAI moves towards a more unified AI experience with GPT-5.
In a series of posts on X (formerly Twitter), Altman acknowledged that OpenAI’s current model selection process has become too complex and that the company is working toward a system that will seamlessly integrate all AI functionalities into one cohesive experience.
“We want to do a better job of sharing our intended roadmap, and a much better job simplifying our product offerings,” Altman wrote. “We want AI to just work for you; we realize how complicated our model and product offerings have gotten.”
Register for Tekedia Mini-MBA edition 17 (June 9 – Sept 6, 2025) today for early bird discounts. Do annual for access to Blucera.com.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register to become a better CEO or Director with Tekedia CEO & Director Program.
He further added, “We hate the model picker as much as you do and want to return to magic unified intelligence.”
Altman was referring to the current structure of ChatGPT, which requires users to manually choose between different models for various tasks. For instance, users must select GPT-4o when scheduling tasks or opt for o3-mini, which operates faster but is optimized for advanced reasoning. Similar issues exist in competitor chatbots like Google’s Gemini, which offers multiple models such as Flash, Pro, and Pro with Deep Research.
The rollout of GPT-4.5 will mark a turning point in OpenAI’s approach, as it will be the final “non-chain-of-thought model” before the company consolidates its AI lineup. Altman explained that OpenAI is working to merge the o-series models, such as o3-mini, with the more advanced GPT-series. This means that rather than forcing users to select the right AI model for each task, OpenAI will build a system capable of making those decisions automatically.
Following the release of GPT-4.5, Altman confirmed that OpenAI “will no longer ship o3 as a standalone model.” Instead, the model’s functionalities will be integrated directly into GPT-5. This change signifies a major shift in OpenAI’s product philosophy, aiming to remove complexity while ensuring users always have access to the most advanced AI capabilities without having to manually switch between models.
Competition From China Driving OpenAI’s Fast-Paced Roadmap
The urgency behind OpenAI’s roadmap has been partly driven by increasing competition from Chinese AI firms, particularly DeepSeek, a startup that has recently gained attention for its rapid advancements in AI technology. In January, DeepSeek unveiled an open-source AI model that impressed users with its reasoning capabilities and efficiency, prompting concerns among U.S. tech giants about China’s growing influence in the AI race. DeepSeek’s entry into the market even triggered a sell-off in global tech stocks, highlighting the disruptive potential of Chinese AI firms.
Altman’s roadmap announcement comes at a time when companies like Baidu, Alibaba, and ByteDance are aggressively developing AI models to rival OpenAI’s ChatGPT. The rapid cost reduction in AI inference, as noted by Baidu CEO Robin Li at the World Governments Summit in Dubai, has added pressure on OpenAI to stay ahead of the curve.
“The inference cost [of foundation models] basically can be reduced by more than 90% over 12 months,” Li stated. “If you can reduce the cost by a certain percentage, then that means your productivity increases by that kind of percentage. I think that’s pretty much the nature of innovation.”
GPT-5 to Offer Free Unlimited Access for Standard Users
One of the most significant revelations from Altman’s announcement was the decision to expand access to OpenAI’s most advanced models. Currently, free users of ChatGPT have limited interactions with the latest AI models, with access to GPT-4o being restricted. However, with the release of GPT-5, OpenAI plans to eliminate these restrictions, granting free-tier users unlimited access to the model at “standard intelligence settings.”
“The free tier of ChatGPT will get unlimited chat access to GPT-5 at the standard intelligence setting (!!),” Altman wrote. He clarified that access would still be “subject to abuse thresholds,” meaning that OpenAI will likely implement safeguards to prevent excessive or exploitative use.
Meanwhile, OpenAI’s paid subscription tiers, including Plus and Pro, will continue to exist, with Altman stating that higher-tier users will receive enhanced capabilities. However, he did not specify what features will differentiate these subscription levels, leaving room for speculation on whether Pro users will get access to more advanced functionalities, faster processing, or larger context windows.
GPT-5 to Feature Advanced Multimodal Capabilities
In addition to simplifying its product lineup, OpenAI is making major improvements to GPT-5’s functionality. Altman confirmed that the upcoming model would feature a much larger context window, allowing it to process and retain significantly more information in long conversations. GPT-5 is also expected to introduce more agentic capabilities, meaning it will be better equipped to take on tasks that require reasoning, decision-making, and minimal user intervention.
These improvements align with OpenAI’s broader goal of developing AI systems that can function as intelligent personal assistants rather than just chatbots.
“These models will incorporate voice, canvas, search, deep research, and more,” Altman stated.
The mention of “canvas” suggests that OpenAI could be working on an interactive workspace feature, potentially allowing users to visualize ideas, create charts, or manipulate data within the ChatGPT environment.
Expected Release Timeline for GPT-4.5 and GPT-5
While Altman did not provide a specific release date for the new models, he hinted that users wouldn’t have to wait too long. In response to a question on X about when the models would be available, he said they would be launched “in a matter of weeks or months.”
This suggests that GPT-4.5 could arrive within the first half of 2025, with GPT-5 following soon after. Given OpenAI’s need to maintain its competitive edge against emerging rivals, the company may be accelerating its release schedule to ensure it remains the leader in AI development.