- By Prateek Levi
- Thu, 15 May 2025 03:27 PM (IST)
- Source:JND
OpenAI GPT-4.1 Launched: OpenAI has officially launched its next big upgrade—GPT-4.1 and its family of models—for ChatGPT users. The lineup includes GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano, each bringing some serious improvements across coding, instruction following, and long-form comprehension. In a blog post announcing the rollout, OpenAI wrote, “These models outperform GPT-4o and GPT-4o mini across the board, with major gains in coding and instruction following.”
But there’s a catch: the new models aren’t for everyone. Only paying ChatGPT users—those on the Plus, Pro, or Team plans—will get access for now. OpenAI confirmed the rollout in a post on X (formerly Twitter) on May 14, noting that GPT-4.1 is now live on ChatGPT. The API version of the 4.1 family actually launched a month ago, giving developers early access to the full, mini, and nano variants. Now, those same models are making their way into ChatGPT itself, with free users still limited to GPT-4o.
What’s New with GPT-4.1?
The biggest leap comes in performance. OpenAI claims GPT-4.1 leaves GPT-4o in the dust, especially in software engineering and instruction-heavy tasks. It’s also built with a 1 million token context window, meaning it can remember and process more data at once. Another key upgrade? The knowledge cutoff jumps to June 2024, compared to GPT-4o’s October 2023.
In terms of benchmark results, OpenAI reports that GPT-4.1 shows a 21% absolute improvement over GPT-4o in software engineering tasks and a 10.5% boost in instruction following. The model also now handles multi-turn conversations more effectively—making it more useful for real-world use cases like customer support, coding, and writing tools. As OpenAI puts it, “While benchmarks provide valuable insights, we trained these models with a focus on real-world utility. Close collaboration and partnership with the developer community enabled us to optimise these models for the tasks that matter most to their applications.”
Mini and Nano: Cheaper, Faster Options
The mini and nano versions aren’t just watered-down clones. OpenAI designed them to bring speed and affordability without totally sacrificing performance. GPT-4.1 mini reportedly cuts latency by almost half while costing 83% less than GPT-4o. And GPT-4.1 nano? It’s OpenAI’s lightest, fastest, and cheapest model yet—great for simple tasks like autocomplete or text classification.
As OpenAI explains, “These models push performance forward at every point on the latency curve.”
Availability and Pricing
For now, only paying ChatGPT users—Plus, Pro, and Team—get to try out GPT-4.1 inside the app. Free users will continue using GPT-4o, although OpenAI says that model will slowly gain some of 4.1’s improvements over time.
Developers can also access GPT-4.1 through the API, where it’s positioned as a faster and more cost-effective option than older models. Pricing starts at $2 per million tokens for GPT-4.1, while nano begins at just $0.10 per million tokens. OpenAI has also increased prompt caching discounts to 75%, making repeated prompts more affordable.
This launch also marks a shift in OpenAI’s product lineup. The company has already phased out GPT-4.5 Preview as of April 14, 2025, and discontinued GPT-4, which had powered ChatGPT Plus since March 2023. GPT-4.1 isn’t technically replacing GPT-4o inside ChatGPT, but many of its capabilities are being folded into the GPT-4o experience. For those wanting the absolute best performance, though, direct access to GPT-4.1 through an API or a ChatGPT subscription is now the go-to option.