Why is Gemini incapable of reading text in images?


With everyone shilling the hell out of Gemini lately, I figured it was time to give it another try.

For context, I'm using Pro with Thinking.

I mostly use AI for basic tasks. Research, general searching, and often for sending a screenshot and having it write out the text in the image. It's this point where Gemini seems to totally fail.

I sent it a screenshot of a training log. The text is clear, and easy to read. It outright refused to get a single word from it.

ChatGPT nailed the task perfectly, Grok did a great job with it, Perplexity's default search option also did a good job (the formatting could've been better, though.)

I appreciate this might be a niche use, but I'm often left disappointed with Gemini, and still failing to understand the hype for the most part. Is it just my use case that it's terrible at, or am I missing something?

Leave a Reply