OpenAI has announced that ChatGPT users can now access GPT-4.1 and GPT-4.1 mini AI models. The good news is that GPT-4.1 mini is available for free users.

Techcrunch reports that the new AI models had originally been launched in April, but were only available through its API for developers. OpenAI says that GPT-4.1 has been designed as a specialized model that excels at coding tasks & instruction following. It also said that the new AI model is faster and better than o3 and o4-mini. The new model is available for ChatGPT Plus, Pro, and Team users. Subscribers can select GPT-4.1 in the "more models" dropdown.

GPT-4.1 mini is a fast and efficient small model that offers various improvements over GPT-4o mini. It can handle various tasks such as instruction-following, coding, and overall intelligence. The release notes published by OpenAI confirms that GPT 4.1 mini has replaced the previous 4.0 mini. It is also available for free users as the fallback model, i.e. once you exhaust your GPT-4o usage limit, ChatGPT will switch to GPT-4.1 mini. The rate limits for the models remain unchanged.

OpenAI had been criticized for not publishing a safety report when it launched GPT-4.1 last month. The company seems to have learnt its lesson, and has announced a Safety Evaluations Hub, a special section on its website which contains detailed results about its AI models, and safety evaluations.

According to Johannes Heidecke, Head of Systems at OpenAI, GPT-4.1 builds on the safety work and mitigations that were developed for GPT-4o, and that GPT-4.1 performs at parity with GPT-4o. He highlighted that no new safety risks had been introduced. GPT-4.1 doesn’t introduce new modalities or ways of interacting with the AI, and that it doesn’t surpass o3 in intelligence.

AI services have been adding support for coding tools rapidly, Microsoft has official GitHub Copilot, Google's Gemini can analyze GitHub projects. OpenAI has reportedly reached an agreement to acquire AI-powered coding startup, Windsurf, for $3 Billion.