This website uses Cookies. Click Accept to agree to our website's cookie use as described in our Privacy Policy. Click Preferences to customize your cookie settings.
Accept
Reject
Preferences
Log in to ask a question
Get hands-on experience with 20+ free Google Cloud products and $300 in free credit for new customers.
Log in to ask a question

Wrong input token count for image prompt in gemini-flash-2.0 API

Posted on 02-09-2025 05:17 AM
buzzingai
New Member
Reply posted on --/--/---- --:-- AM

As per this doc, the image in the prompt to gemini API is counted as a fixed 258 tokens. I verified that the same is true for gemini-1.5-flash, gemini-1.5-flash-8b and gemini-2.0-pro-exp-02-05 models. But for gemini 2.0-flash model, I am getting the prompt_token_count as high as 2300 tokens, which is a 9x increase in token count.

I even verified this by sending the image with empty text prompt so that i am sure that text prompt is not taking up any tokens. The exact same code results in prompt_token_count of ~300 for other models.

Anyone else facing this issue ?

 

0 0 31
Topic Labels
0 Likes
0 REPLIES 0