We’re bringing a new dimension to our AI interactions by enabling the GPT-4 Vision model within Straico. This feature allows you to use images as a layer of context in your conversations with our AI, expanding the scope of your queries and the insights you can gain.
Big shoutout to Johan for his awesome work! (He never checks Discord lol)
Getting Started with GPT-4 Vision:
- Within Straico, switch to the OpenAI: GPT-4 vision model.
- Click the ‘+’ adjacent to the chat input box, and select the image tab.
- We support multiple images, you can upload your images individually, in the same manner, you’d add attachments.
- Begin your dialogue with the AI, now with your images as part of the conversation.
Potential Use Cases:
This feature is designed for those moments when text alone can’t capture the full story. From analyzing design elements to translating visual data into actionable insights, we believe this tool will add significant value to your workflow.
- I asked vision to pick a winner in a freestyle fight between Kirby, Elon Musk and Leon from Resident Evil, guess who won: https://platform.straico.com/share/chat/6577c4255f4f4a52c1646ca0?fpr=arturo-promptrack21
- I asked vision to describe a screenshot of our Discord Server: https://platform.straico.com/share/chat/6577c4534db36052c089aeb8?fpr=arturo-promptrack21
About the coin cost per image
The coin cost of a given image is determined by its size.
Images are first scaled to fit within a 2048 x 2048 square, maintaining their aspect ratio. Then, they are scaled such that the shortest side of the image is 768px long. Finally, we count how many 512px squares the image consists of. Each of those squares costs 47 coins. Another 23 coins are always added to the final total.
Yes, the process is somewhat complex, but this is how OpenAI calculates the token usage for vision tasks. Nonetheless, the following benchmarks provide a good point of reference:
- The cost for a 2048×2048 image: 210 coins
- The cost for a 512×512 image: 20 coins
(This info is also available in straico.com/multimodel)
A Collaborative Approach to Innovation:
As we roll out this feature, your engagement is invaluable. Your explorations and feedback are the compass that guides our improvements and helps us refine the user experience.
We’re not announcing this to the broader public just yet—it’s an exclusive opportunity for our community to shape the official release.
Feel free to start using this feature and let us know your thoughts.
