Hello GPT-4o

🌈 Abstract

The article discusses the announcement of a new AI model called GPT-4o by OpenAI. GPT-4o is described as a multi-modal model that can handle text, images, and audio, with improved capabilities compared to previous versions of GPT. The article highlights several key features of GPT-4o, including:

• Ability to act as a live interpreter between people speaking different languages, with improved control over voice and intonation
• Advancements in image generation, particularly in the areas of text output and maintaining consistent characters across prompts
• Increased vocabulary size, resulting in more efficient handling of non-English languages
• Reduced pricing compared to GPT-4 Turbo, with the model being made available to free ChatGPT users for the first time

The article also mentions upcoming support for GPT-4o's new audio and video capabilities, which were hinted at during the launch presentation.

🙋 Q&A

[01] GPT-4o Announcement and Features

1. What are the key new features of GPT-4o compared to previous GPT models?

• GPT-4o is a multi-modal model that can handle text, images, and audio
• It can act as a live interpreter between people speaking different languages, with improved control over voice and intonation
• It has made advancements in image generation, particularly in the areas of text output and maintaining consistent characters across prompts
• It has a larger vocabulary size, resulting in more efficient handling of non-English languages
• It is priced 50% lower than GPT-4 Turbo, and will be made available to free ChatGPT users for the first time

2. What are the upcoming capabilities related to audio and video that were hinted at in the announcement? The article mentions that OpenAI plans to launch support for GPT-4o's new audio and video capabilities to a small group of trusted partners in the API in the coming weeks, but does not provide further details on what these capabilities entail.

[02] Pricing and Availability

1. How does the pricing of GPT-4o compare to previous GPT models? The article states that GPT-4o is priced at $5/million input tokens and $15/million output tokens, which is a 50% reduction compared to GPT-4 Turbo. In comparison, GPT-3.5 is priced at $0.50/million input tokens and $1.50/million output tokens.

2. Who will have access to GPT-4o? The article mentions that GPT-4o will be made available to free ChatGPT users, which is the first time OpenAI has made their "best" model available to non-paying customers.