BIG ChatGPT Update

Hi all,

If you haven’t heard the recent news in AI, then I’ll break it to you.

OpenAI has introduced a new line of AI models called OpenAI o1, designed to excel at complex reasoning tasks in science, coding, and math.

I’ll briefly break it down, so that you can be updated in as little time as possible.

A short summary of o1:

  1. Enhanced Reasoning: The models spend more time "thinking" before responding, improving problem-solving capabilities.

  2. Impressive Performance: The upcoming model update performs similarly to PhD students in physics, chemistry, and biology benchmarks. It scored 83% on International Mathematics Olympiad problems and reached the 89th percentile in coding competitions.

  3. Safety Improvements: New safety training approaches leverage the models' reasoning abilities to better adhere to guidelines.

  4. Two Models Available:

    • o1-preview: Main model with broad capabilities

    • o1-mini: Faster, cheaper model optimized for coding tasks

You might now be thinking to yourself, "What is the difference between the OpenAI models?"

Simply put, this is the difference between the most popular models from OpenAI:

  • GPT-3.5 Turbo: Fast and cheap model for simple tasks.

  • GPT-4 Turbo & GPT-4: High-intelligence models. Still relatively fast and more expensive, good for more complex tasks.

  • GPT-4o: High-intelligence model for complex, multi-step tasks.

  • GPT-4o mini: Affordable and intelligent model for fast, lightweight tasks.

  • o1-preview (OpenAI o1): The slowest and most expensive model for the most complex tasks that require the highest level of reasoning.

  • o1-mini (OpenAI o1): Faster and less expensive model for complex tasks that are lightweight and still require high-level reasoning.

  • DALL·E: A model that can generate and edit images.

In a future newsletter, I will break down all free AI platforms that you can use for day-to-day use, along with a comparison of when to use each one and which LLMs to use for each scenario.

Can I use the new model, and if so, when?

  • ChatGPT Plus and Team users have immediate access.

  • ChatGPT Enterprise and Edu users will have access at the beginning of this week (Week 38).

  • API access is available for qualified developers.

  • For free users, there are plans to bring o1-mini to all ChatGPT Free users.

A Personal Update - Agentflow

We will also begin to use the o1 model for the AI chatbot solutions at Agentflow when it is integrated natively within the platform. This will drastically improve the quality of the answers provided by the chatbots we develop.

If you want to find out if we can help you handle customer service through an AI employee on your website that works 24/7, 365 days of the year. Then book a Free demo here 👇

That’s all for this update. I hope that you found it short and concise (which is the goal for the newsletter).

Back soon,

Axel

How was todays letter?

Reply to the email to share feedback, it is greatly appreciated.