Tech Desk 
Sep 15, 2024
OpenAI has been both the developer of some of the best artificial intelligence (AI) and sort of a buzzword among AI freaks, and that has been well justified in view of the slew of OpenAI-made product launches following a short interlude after the release of ChatGPT.
While debunking the prevailing rumours about AI models dubbed “Strawberry” and "GPT-5", OpenAI earlier this week startled the world by announcing new large language models (LLMs) called o1, a new addition to the series whose theme is “reasoning”.
The new o1 models from OpenAI are designed to offer more efficient and precise responses for queries involving science, technology, engineering and math (STEM) fields.
Read more: Anthropic launches Claude Enterprise plan to challenge OpenAI's dominance
OpenAI's new o1 models: How can developers make the most
Exclusively available to ChatGPT Plus users and developers with OpenAI’s paid application programming interface(API), o1-preview and o1-mini are the two models OpenAI offered with the new o1 lineup of LLMs.
Be it for testing them as the backend of existing third-party apps and services, or developing new ones on their bases, developers can make the most of these OpenAI's new o1 models.
But like any other byproduct of technology, the discouraging point is that o1 models are limited to text inputs and outputs, which is contrary to what GPT-4o was capable of.
Apart from this, another woe is that developers cannot access external information or analyse files, owing to the restricted feeding of data they were trained on before October 2023.
Lastly, OpenAI's o1-preview model is more expensive than other OpenAI models, however, o1-mini could be considered to be affordable.
