OpenAI Launches New Coding-Focused GPT-4.1 Models
OpenAI today announced the launch of three new GPT models that are available through the OpenAI API. The models, called GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano, come with major improvements in coding, instruction following, and handling long context, as per OpenAI.
The GPT-4.1 models have shown superior performance compared to GPT-4o and GPT-4o mini, especially in coding tasks. They now support up to one million tokens of context, enhancing their long-context comprehension abilities.
Performance Improvements
On SWE-bench Verified, GPT-4.1 demonstrated a 21.4 percent improvement in coding over GPT-4o and a 26.6 percent improvement over GPT-4.5. The slimmed-down GPT-4.1 mini matches or even surpasses GPT-4o with reduced latency and an 83 percent lower cost. Meanwhile, GPT-4.1 nano stands out as OpenAI's fastest and most cost-effective model, perfect for tasks such as classification and autocompletion.
Feature Incorporation
OpenAI has already integrated many of the GPT-4.1 improvements into the ChatGPT version of GPT-4o, with plans for further enhancements in the future. All GPT-4.1 models are up to date until June 2024, ensuring they are current on events preceding that date.
Deprecation of GPT-4.5
With the introduction of GPT-4.1, OpenAI has deprecated GPT-4.5 in the API, as the new models offer similar functionality at a reduced cost.
For more information on the new OpenAI GPT-4.1 models and their pricing, you can visit the official OpenAI website.
