After US Approval, OpenAI Released New GPT-5.6 Models

Following the approval of the US government, artificial intelligence giant OpenAI made its most powerful models, the GPT-5.6 series, available to the general public as of Thursday. The company's new generation AI models had raised national security concerns due to their unique capabilities in detecting coding errors and finding cybersecurity vulnerabilities. The Washington administration closely monitored the rollout process of these models to assess potential risks. As part of the process, OpenAI shared a preview version of the GPT-5.6 model with a trusted group of local partners at the request of US officials in late June. Following technical tests and intensive discussions with government officials, the company finally got the green light for a broad rollout.

The GPT-5.6 series will be available in three different tiers carefully designed to meet various user needs. The first of these, 'Sol', stands out as the company's most advanced model, featuring flagship capabilities. The 'Terra' model, optimized for mid-level users and daily workloads, aims to strike a balance between performance and accessibility. Lastly, the model named 'Luna' is positioned as a fast and economical option, aiming to minimize costs while providing high speed. OpenAI announced that it is gradually expanding global preview access to all these models. In its announcement made via the social media platform X, the company stated that details of the new features to be offered to users would be shared in the coming days.

According to a report by the US news agency Axios, based on sources, the Trump administration officially permitted the company to conduct a broad rollout. Technical reviews conducted by the company with government officials and strategic meetings regarding potential cybersecurity risks were decisive in making this decision. This move shows a significant parallel with the steps taken by Anthropic, another major industry player and the company behind the Claude chatbot. Anthropic also announced last week that it started to restore access to its most powerful AI models, Fable 5 and Mythos 5, following the US government's removal of geographic restrictions. Thus, the competition in the AI sector continues unabated, accompanied by the guided regulatory policies of the US government.

The ability of AI models to detect vulnerabilities in code continues to be debated as an important issue for national security. In line with an executive order from the White House, the government is currently in the process of preparing clear criteria to determine which models will be subject to new security restrictions. The administration of President Donald Trump had previously indicated, through earlier statements, a tendency to relax the rules applied to AI companies in order to win the technology race against China. However, in its statement in June, OpenAI argued that government intervention in model access processes is not a sustainable model in the long run. The company emphasized that this process, which it criticized for keeping the best tools away from users, businesses, and those in need, negatively impacts the user experience.

In this fiercely competitive environment in the tech world, giants like OpenAI, Anthropic, and Google are constantly updating their cost strategies. When the GPT-5.6 series becomes widely accessible, the pricing of the mid-level 'Terra' model is reported to be half that of its predecessor, GPT-5.5. The primary goal of this aggressive pricing strategy is to retain customers in the market and gain a competitive advantage against rivals. The scale of commercial competition is not limited to price wars; OpenAI and Anthropic have submitted confidential Initial Public Offering (IPO) documents to US regulators, with each company planning to go public, targeting a valuation of around 1 trillion dollars. The fact that both companies are preparing to go public significantly raises the economic stakes of this AI arms race between them on a global scale.