OpenAI has launched its next generation of LLMs—the GPT-4.1 family, comprising GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano. These new models surpass the capabilities of GPT-4o and GPT-4o mini, showcasing significant advancements in coding, instruction following, and long context comprehension, all while offering lower costs and reduced latency.
The new models are available immediately via the API, with GPT-4.1 priced at $2.00 per 1 million input tokens and $8.00 per 1 million output tokens. GPT-4.1 mini costs $0.40 and $1.60, respectively, and GPT-4.1 nano is priced at $0.10 and $0.40. OpenAI is also increasing the discount for repeated context use to 75% for these new models.
“GPT‑4.1 is a significant step forward in the practical application of AI,” OpenAI said in its announcement. The models feature an expanded context window of up to 1 million tokens and incorporate knowledge up to June 2024.
In coding proficiency, GPT-4.1 achieved a 54.6% score on the SWE-bench Verified benchmark, an improvement of 21.4 percentage points over GPT‑4o. Its ability to follow instructions also saw gains, scoring 38.3% on the Scale MultiChallenge benchmark, a 10.5 percentage point increase over GPT‑4o. For understanding long sequences of information, GPT-4.1 reached a new high of 72.0% on the Video-MME benchmark.
The GPT-4.1 family also emphasises efficiency. GPT-4.1 mini “matches or exceeds GPT‑4o in intelligence evaluations while reducing latency by nearly half and reducing cost by 83%”, OpenAI further said.
In the 4.1 family, GPT-4.1 nano is the fastest and cheapest model available for tasks that need quick responses. It has a 1 million token context window and performs well in tasks like classification and autocompletion.
These advancements are expected to facilitate the development of more capable AI agents that can independently handle tasks.
“Developers can now build agents that are more useful and reliable at real-world software engineering, extracting insights from large documents, resolving customer requests with minimal hand-holding, and other complex tasks,” the company stated.
It is important to note that GPT-4.1 will be accessible solely through the API. OpenAI clarified that many of the improvements in instruction following, coding, and intelligence have been gradually incorporated into the latest version of GPT‑4o within ChatGPT.
OpenAI also announced the upcoming discontinuation of GPT-4.5 Preview in the API, scheduled for July 14, 2025. This decision was made because “GPT‑4.1 offers improved or similar performance on many key capabilities at much lower cost and latency”.
Early testing by various companies has yielded positive results. According to the company, Windsurf reported that GPT-4.1 scores “60% higher than GPT‑4o on Windsurf’s internal coding benchmark”, leading to “faster iteration and smoother workflows”.
Moreover, Qodo found that GPT-4.1 produced better suggestions in 55% of cases for code reviews. BlueJ noted a 53% more accurate performance on complex tax scenarios, while Hex observed nearly two times more improvement on challenging SQL evaluations. Thomson Reuters experienced a 17% improvement in multi-document review accuracy, while Carlyle reported 50% better retrieval from large, data-rich documents.
The post OpenAI Unveils GPT-4.1 Models in API Specifically for Coding appeared first on Analytics India Magazine.