OpenAI has launched its new models, GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano, marking notable enhancements over the previous GPT-4o, especially in programming, instruction adherence, and long context processing. Each model offers varying capacities and cost levels, supporting up to 1 million tokens. The flagship GPT-4.1 shows a 21.4% improvement in coding capabilities, while mini and nano variants provide rapid and cost-effective solutions. These developments are tailored for developers needing efficient performance in producing autonomous agents and processing large datasets.
OpenAI has introduced the GPT-4.1 family of models, which significantly outperforms GPT-4o, especially in programming and instruction-following capabilities.
The GPT-4.1 model scores 54.6% on SWE-bench Verified, showing a 21.4% improvement over GPT-4o, underscoring its enhanced coding abilities.
With a new capacity to process 1 million tokens and an improved score of 72% on Video-MME, GPT-4.1 excels at long context handling.
The introduction of GPT-4.1 mini and nano variants allows developers to access powerful solutions for low-latency applications at reduced costs.
Collection
[
|
...
]