OpenAI today released a new series of advanced reasoning models called “o1” for ChatGPT. These models share the codename “Strawberry” and represent a paradigm shift in artificial intelligence. OpenAI o1 models match human experts in science-related fields like math, physics, chemistry, and coding. OpenAI claims that its new ChatGPT o1 models are on par with PhD students, which is quite significant.
The most interesting thing is that two new o1 models – ChatGPT o1-preview and ChatGPT o1-mini – are already available for ChatGPT Plus subscribers. That said, there is a weekly rate limit of 30 messages for o1-preview and 50 for o1-mini. So let’s see how you can test the ChatGPT o1 model.
How to use the ChatGPT o1 reasoning model
- Visit chatgpt.com (visit) and log in with your account.
- Now upgrade to the ChatGPT Plus plan which costs $20 per month.
- Once you upgrade to the paid plan, click the drop-down menu in the upper left corner and select “o1-preview” or “o1-mini”.
- Now you can start chatting with ChatGPT o1 models. For example, I asked a series of reasoning questions and it solved all of them without any problems.
OpenAI o1 models can tackle complex reasoning questions and solve much more difficult problems. In my tests so far, ChatGPT o1 has answered all reasoning questions correctly I gave it a try. Previously, even the most advanced LLMs like GPT-4o and Claude 3.5 Sonnet failed to solve common sense reasoning problems and logical questions.
OpenAI claims that the o1 models were trained to perform a chain of thought reasoning through reinforcement learning. This is why it takes some time to “think” before generating a response.
Keep in mind that you can’t use o1 models to analyze documents, images, or browse the web. Currently, it only supports text input. OpenAI says that multimodal input capability will be available later. As for free ChatGPT users, the ChatGPT o1-mini model should be available soon. Anyway, that’s it for us. Go ahead, test OpenAI’s new o1 models, and share your thoughts with us.
Read More:-
