OpenAI rolls out GPT-4.1 series with major upgrades in coding, comprehension

News

1. India’s data boom is just getting started, study finds

2. India’s Digital Future Hinges on Data Quality, NITI Aayog Says

3. India’s Back Offices Become New Frontline in Global Cyber Threats

4. ChatGPT Users Show Lower Brain Activity and Recall, Study Finds

5. GenAI will shrink 'corporate' ranks, Amazon's Jassy says

6. Apex Audit Body Taps AI for Faster, Smarter Audits: Report

OpenAI has released a new GPT-4.1 series, packed with several upgrades to make it better at coding, following instructions, and understanding large amounts of information.

The series includes three versions: GPT-4.1, GPT-4.1 Mini, and GPT-4.1 Nano. Each one is made for different developer needs, with a focus on speed, lower costs, and practical use.

A major improvement in GPT-4.1 is its ability to handle up to one million tokens at a time, which is a big jump from the earlier GPT-4o that could only handle 128,000 tokens.

Tokens are like building blocks of a sentence that AI uses to read, understand, and generate text.

With this upgrade, the new models can now manage entire codebases, long documents, or complex instructions more easily.

Tests show that GPT-4.1 performs better than previous models. It is 21% better at coding than GPT-4o and 27% better than GPT-4.5.

It also works 40% faster and is 26% cheaper. The cost is $2 for every million input tokens and $8 for every million output tokens.

OpenAI said GPT-4.1 scored 54.6% on the SWE-bench Verified test, slightly behind models from Google and Anthropic but at a much lower price.

The SWE-bench Verified test measures how well an AI can read software problems and suggest helpful fixes, like a smart coding assistant.

In a live demo, GPT-4.1 was used to build a flashcard app, move through folders, and write code that worked without errors. It also did well in tasks like understanding code files and running tests.

OpenAI CEO Sam Altman said, “These models are great at coding, instruction following, and long context. Benchmarks are strong, but we focused on real-world applications, and developers seem very happy.”

Each model in the GPT-4.1 series has a different use, with the standard version offering all features.

GPT-4.1 Mini is cheaper and faster. GPT-4.1 Nano is the smallest and best for simple tasks like sorting data or finishing sentences.

According to OpenAI, the latest launch is part of a larger plan to build an “agentic software engineer”—an AI that can help with all parts of software development, from writing and testing code to fixing bugs and writing instructions.

ChatGPT will also remove GPT-4 by 30 April, stop supporting GPT-4.5 in its API by 14 July to make room for new models.