Introduction

Today, we’re launching three new models in the API: GPT‑4.1, GPT‑4.1 mini, and GPT‑4.1 nano. These models outperform GPT‑4o and GPT‑4o mini across the board, with major gains in coding and instruction following. They also have larger context windows—supporting up to 1 million tokens of context—and are able to better use that context with improved long-context comprehension. They feature a refreshed knowledge cutoff of June 2024. GPT‑4.1 excels at the following industry standard measures: Coding: GPT‑4.1 scores 54.6% on SWE-bench Verified, improving by 21.4%abs over GPT‑4o and 26.6%abs over GPT‑4.5—making it a leading model for coding. Instruction following: On Scale’s MultiChallenge⁠(opens in a new window) benchmark, a measure of instruction following ability, GPT‑4.1 scores 38.3%, a 10.5%abs increase over GPT‑4o. Long context: On Video-MME⁠(opens in a new window), a benchmark for multimodal long context understanding, GPT‑4.1 sets a new state-of-the-art result—scoring 72.0% on the long, no subtitles category, a 6.7%abs improvement over GPT‑4o.

Benchmarking

Training

Authors

+0

Research leads Ananya Kumar, Jiahui Yu, John Hallman, Michelle Pokrass Research core contributors Adam Goucher, Adi Ganesh, Bowen Cheng, Brandon McKinzie, Brian Zhang, Chris Koch, Colin Wei, David Medina, Edmund Wong, Erin Kavanaugh, Florent Bekerman, Haitang Hu, Hongyu Ren, Ishaan Singal, Jamie Kiros, Jason Ai, Ji Lin, Jonathan Chien, Josh McGrath, Julian Lee, Julie Wang, Kevin Lu, Kristian Georgiev, Kyle Luther, Li Jing, Max Schwarzer, Miguel Castro, Nitish Keskar, Rapha Gontijo Lopes, Shengjia Zhao, Sully Chen, Suvansh Sanjeev, Taylor Gordon, Ted Sanders, Wenda Zhou, Yang Song, Yujia Xie, Yujia Jin, Zhishuai Zhang