OpenAI unveils GPT-5 model, featuring improved coding and problem-solving chops

OpenAI on Thursday unveiled its highly anticipated GPT-5, a powerful multi-modal AI model featuring major advancements in problem-solving and coding.

The new flagship model was announced during a Thursday morning livestream. Unlike previous releases that were limited to paid subscribers, GPT-5 will be available to free-tier ChatGPT users as well, OpenAI said.

OpenAI CEO Sam Altman described interacting with the new model as akin to conversing with a PhD-level expert, noting that while GPT-3 was comparable to a high school student and GPT-4 to a college student, “With GPT-5 you get an entire team of PhD experts in your pocket, ready to help you,” Altman said.

During the announcement, OpenAI researchers emphasized that GPT-5 was designed to be more reliable and accurate, with fewer hallucinations.

GPT-5 offers improvements in reasoning, higher-quality code generation, greater autonomy with reduced need for user input, and seamless integration with platforms like ChatGPT and Google’s Gmail and Calendar apps.

Earlier language models relied only on pretraining to generate responses. GPT-5, like recent inference-based models, can also incorporate new data from user prompts in real time (a method known as “test-time computing”). This release also unifies OpenAI’s model naming, replacing names like o1 and o4-mini with the GPT-5 family, signaling a shift to models that combine both pretraining and inference.

For many, the key question is whether the leap from GPT‑4 to GPT-5 will prove as dramatic as the jump from GPT‑3 to GPT‑4. Independent testing will shed light on that.

OpenAI also recently introduced two “open-weight” models—gpt-oss-120b and gpt-oss-20b—which are freely available and modifiable by developers. This move marked a rare shift from its typically closed model strategy. However, with GPT-5, OpenAI returns to its more traditional “closed” approach.

In addition, the company announced a partnership with the U.S. federal government to provide executive branch agencies access to its enterprise-grade chatbot. Through a landmark agreement with the General Services Administration (GSA), ChatGPT Enterprise will be made available to agencies for just $1 per agency for one year. OpenAI has assured that it will not use government data to train its AI models.

https://www.fastcompany.com/91382136/openai-unveils-gpt-5-model?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss