OpenAI Launches GPT-5 Free to All ChatGPT Users
OpenAI has introduced GPT-5, along with three variants: GPT-5 Pro, GPT-5 mini, and GPT-5 nano, acclaimed as their "best AI system yet." This release grants free access to some models for ChatGPT users across all tiers. Notably, GPT-5 brings enhancements in handling sensitive requests, dubbed "safe completions," along with better coding capabilities and fewer confabulations.
This debut represents a significant milestone as free users gain access to a simulated reasoning AI model for the first time. By breaking down problems into steps, this technique enhances the accuracy of responses to logical or analytical questions.
GPT-5 consolidates OpenAI's diverse AI functions into a unified platform, featuring a smart model for typical inquiries, a deep reasoning model known as "GPT-5 thinking," and a real-time router to determine the best approach based on various factors. Like its predecessor, GPT-5 is a multimodal system capable of handling images, voice, and text interactions.
The deployment commences today, impacting ChatGPT's 700 million weekly active users, with usage limits differing by subscription tier. Pro subscribers enjoy unlimited access to GPT-5 and its Pro variant, while Plus users experience elevated usage limits compared to free users. GPT-5 Pro supersedes o3-pro for those eligible subscribers.
Despite previous trends of diminishing advancements with each new AI model, GPT-5 still makes notable progress from previous iterations such as GPT-4o. While the upgrade might feel incremental, OpenAI's use of the "GPT-5" name capitalizes on brand recognition amid fierce industry competition.
In coding benchmarks, GPT-5 achieves 74.9% on SWE-bench Verified and 88% on Aider Polyglot benchmarks, presenting the "strongest coding model yet" from OpenAI. In health-related inquiries, GPT-5 scores 46.2% on HealthBench Hard, assisting users in understanding medical outcomes but with the caution not to replace professional advice.
Additional updates include improved interface customization, personality presets, and new integrations with Gmail and Google services for Pro users. OpenAI has also moved towards "safe completions" as a refined approach to response censorship, aiming to provide helpful replies within safety boundaries.
Developers can access GPT-5 in three API versions, with significant expansions in context window size and flexible pricing models. New features for developers include "free-form function calling" and "reasoning effort control." This rollout comes in response to competitive pressures from other AI players like Google's Gemini and Anthropic's Claude family.
