OpenAI officially launched GPT‑5 on August 7, 2025.

The new model is now accessible to all ChatGPT users—including free, Plus and Pro tiers—as well as via the API .

GPT‑5 unifies multiple model strengths into a single architecture, combining fast processing, advanced reasoning and multimodal input handling.

It supports text, image, audio and video inputs, and automates task routing through a real‑time model selection mechanism .

The model demonstrates enhanced capabilities in coding, reasoning, writing, health and finance tasks.

Benchmark results place GPT‑5 at the top of SWE‑Bench for software engineering tasks, Aider Polyglot for multilingual coding, and other reasoning and health‑related evaluations .

OpenAI has implemented measures to reduce hallucinations and improve reliability, including structured reasoning and test‑time compute to better handle complex or open‑ended queries .

GPT‑5 is available in multiple variants: the standard model, GPT‑5‑mini, and GPT‑5‑nano, optimized for cost‑efficient and high‑speed API use; Pro subscribers also receive access to GPT‑5‑pro and GPT‑5‑thinking for deeper processing and higher usage limits .

Context window capacity has been significantly extended, with support for up to 256,000 tokens—or potentially up to one million in certain implementations—facilitating long‑form interaction and content processing .

GPT‑5 integrates with external tools and services, including Gmail, Google Calendar, and voice interfaces, and introduces customizable conversational personalities to adapt the user experience .

The rollout arrives as ChatGPT maintains approximately 700 million weekly active users.

OpenAI is pursuing a valuation near $500 billion, amid continued industry investment in AI infrastructure and competition across the sector .