OpenAI Unveils GPT‑5 With Ph.D.-Level Intelligence, Boosts Writing and Coding Performance

Benzinga

OpenAI unveiled GPT‑5 on Thursday, calling it its smartest, fastest and most useful model yet, according to CNBC. CEO Sam Altman described the new version of ChatGPT as a Ph.D.-level expert, capable of generating more accurate responses, stronger reasoning, and improved articulation. The company said GPT‑5 reduces factual errors by 45% compared to GPT‑4o and enhances performance in coding, health-related questions, and writing. The update also includes fewer hallucinations, toned-down flattery, and limited emoji use. OpenAI will offer ChatGPT-5 free globally as part of its mission to make AI accessible and beneficial to all humanity, according to product head Nick Turley. ChatGPT now serves over 700 million weekly users.

Transcript

00:00It's Benzinga, bringing Wall Street to Main Street.

00:02OpenAI unveiled GPT-5 on Thursday, calling it a major leap forward in coding, reasoning,

00:07and user customization, according to Axios.

00:09Model will power both free and paid chat GPT offerings, and aims to compete directly

00:13with Anthropics Plawn, which is widely considered the top AI for developers.

00:17CEO Sam Altman said GPT-5 enables on-demand software creation, and described the shift

00:22as transformative.

00:23Model combines traditional LLM architecture with advanced reasoning, also reducing hallucinations

00:28and offering customizable personalities.

00:31Early testers praise its ability to debug, run complex background agents, and collaborate

00:35intuitively.

00:36Critics, however, argue GPT-5 lacks the impact needed to tackle global crisis and may worsen

00:41existing problems such as job disruption and deepfakes.

00:45OpenAI's annual revenue run rate now stands at $12 billion, compared to Anthropics' $5 billion

00:49pace.

00:50For all things money, visit Benzinga.com.

Category

Transcript

Be the first to comment

Recommended