OpenAI Announces GPT-5 Release Event, Aims for 10x Intelligence Boost

OpenAI, a prominent artificial intelligence company, has officially announced the upcoming release of GPT-5, its next-generation language model. The company's official account on the social media platform X revealed that a live stream event is scheduled for Thursday morning at 10 AM Pacific Time, which translates to Friday at 1 AM Beijing Time. The clever use of the number "5" in place of the letter "S" in the word "LIVESTREAM" has sparked speculation that the event will focus on the highly anticipated GPT-5 model.

In June, OpenAI's CEO conducted an in-depth interview where they mentioned that GPT-5 is likely to be released this summer. Subsequent reports suggested that OpenAI is planning to launch GPT-5 in August, along with mini and nano versions, which will be accessible to external users through an API. This strategic move aims to make advanced AI capabilities more accessible to a broader range of developers and businesses, fostering innovation and integration of cutting-edge language processing into various applications.

The CEO described GPT-5 as a comprehensive system that integrates a multitude of technological advancements, including enhanced reasoning capabilities. Unlike previous models, GPT-5 is not being released as an independent model but as part of a larger system designed to simplify the product line and merge large language models. This approach aims to create a more powerful system that ultimately achieves the goal of general artificial intelligence (AGI).

OpenAI has been actively releasing new products. Recently, they launched GPT-oss-120b and GPT-oss-20b, two open-weight large models that do not require payment or licensing for commercial use before deployment. While there is no official information indicating that GPT-5 will be open-source, the CEO had previously promised in February that users would be able to access this model for free. The specific details of GPT-5's capabilities are not yet fully disclosed, but it is expected to integrate the foundational models of the GPT series, the reasoning models of the o series, and the full-modal models of the 4o series.

In recent months, there have been numerous updates to large models both domestically and internationally. For instance, Google introduced the general-purpose world model Genie 3, which can generate dynamic worlds in real-time at 24 frames per second and maintain visual consistency for several minutes at 720p resolution. Additionally, Kimi released K2, a model with enhanced coding capabilities, and Zhongxin released GLM-4.5, aiming to achieve breakthroughs in reasoning and agent capabilities.

Analysts believe that the next generation of models, such as GPT-5, will achieve comprehensive development in technology, capabilities, and ecosystem. These models are expected to scale up by 2-3 times, resulting in a nearly 10-fold increase in intelligence. This improvement in the cost-benefit ratio of reasoning will enhance the potential for application profitability. The advancements in logical reasoning, native multimodal capabilities, and memory systems will accelerate the implementation of high-value, complex industry scenarios. As the ability to solve complex tasks improves and the cost-benefit ratio of reasoning enhances, GPT-5 and similar models will become the foundation for the next generation of AI applications. The focus on agents and multimodal capabilities will continue to drive industry advancements and simultaneously boost the demand for reasoning-side computing power.