OpenAI has unveiled a new family of AI models designed to enhance coding capabilities amidst rising competition from tech giants like Google and Anthropic. The models, named GPT 4.1, GPT 4.1 Mini, and GPT 4.1 Nano, outperform previous versions, including GPT-4o and GPT-4.5. With a notable score of 55 percent on the SWE-Bench coding benchmark, the new models demonstrate significant improvements, excelling in complex instruction following and functional code generation. OpenAI's efforts in refining these models reflect the growing significance of AI in software development.
OpenAI is releasing three sizes of models: GPT 4.1, GPT 4.1 Mini, and GPT 4.1 Nano, which excel at coding and complex instruction following.
GPT-4.1 scored 55 percent on SWE-Bench, outperforming other models from OpenAI and indicating a significant improvement in coding capabilities.
"Developers care a lot about coding and we've been improving our model's ability to write functional code," said Michelle Pokrass of OpenAI.
The capacity for AI models to write and edit code has improved significantly, fostering more automated software prototyping and the rise of improved AI agents.
Collection
[
|
...
]