GPT 4 Turbo With Vision Now Available in ChatGPT: What it Means?

JOIN US
Highlights
  • OpenAI has released a new and powerful version of GPT, which can process eight times more data than the previous version.
  • It is trained with information up to December 2023 to achieve better responses based on the latest data.

OpenAI has introduced support for GPT 4 Turbo with Vision in its popular AI chatbot ChatGPT. With this update, users can provide image input to ChatGPT to get responses. The feature is currently available to paid ChatGPT Premium subscribers, which is a paid option. Let’s understand this new GPT model and how it will be helpful for ChatGPT users.

GPT 4 Turbo with Vision Explained

The biggest upgrade with GPT 4 Turbo is the amount of information it can process. The latest version, GPT 4 Turbo with Vision, supports up to 1,28,000 tokens, eight times more capable than GPT 3.5 Turbo, which could support a maximum of 16,385 tokens. Tokens matter because this is essentially the most basic input of data or chunk of text, fed to a large language model (LLM) such as GPT 4 Turbo.

GPT 3.5 Turbo was trained on data available up to September 2021, and the latest GPT 4 Turbo has information up to December 2023, meaning it is more up-to-date with recent events.

For general users, GPT 4 Turbo’s best advantage is its ability to accept input in the form of images and videos. ChatGPT can now respond to users by analyzing multimedia content. According to OpenAI’s blog post, users can provide direct photo/video input or include URLs.

However, OpenAI has warned users that GPT 4 Turbo with Vision has limitations. Although it is the company’s most advanced model, it is still unsuitable for processing and interpreting medical images like CT scans and X-ray reports.

The company further mentions that GPT 4 Turbo with Vision works best with languages that use Latin alphabets. Languages with non-Latin alphabets, such as Korean, Japanese, etc., may not be processed optimally by this model.

Interestingly, OpenAI warns users that the new model struggles with upside-down images and photos with fish-eye effects. For safety reasons, OpenAI has also blocked the usage of GPT 4 Turbo with Vision to clear CAPTCHAs.

GPT 4 Turbo with Vision is currently available to all ChatGPT Plus users. The subscription costs $20 (approx Rs 1660) per month in India.