Analyzing images with GPT-4 Vision

OpenAI just released GPT-4 Vision, a new AI model that can analyze and generate images.

To test how accurate this new model is at analyzing images, I created a Glide template where I can easily upload images, tweak prompts, and run multiple tests.

In general, the results are pretty amazing.

GPT-4 Vision was able to describe all images in incredible detail and provide feedback (and even scores) on how to improve/fix or compare/recommend the specific content in the image.

I ran 28 tests under 11 categories ranging from analyzing meals, receipts, resumes, website designs all the way to providing analysis and feedback on game strategy, car repairs, or even human beauty.

The applications of this technology are endless. However, this model is still in its infancy and (like most LLM tools) is prone to mistakes and hallucinations.

This is why it is important to always double-check all results before making any decisions or deploying any apps into production.

I uploaded a video describing all my results here → [ WATCH VIDEO HERE ]​

As usual, you can get a copy of the final template for the app so you can play with it and learn how it works.

You can grab a copy with a 50% discount → https://grumo.com/getgptvision​

👉 GET A COPY OF THIS TEMPLATE HERE (50% OFF)

Peace. Love. GPT-VISION-Cookies 🤖🍪

Miguel @ Grumo.com

P.S. Want to learn how to build apps with Glide? for more Glide tutorials make sure to subscribe to my YouTube channel → HERE.