Hi. We are using Gemini Omni 1.1 flash via Ai Studio for video generation. However, we are seeing that the output tokens are being calculated incorrectly - irrespective of the duration of the video and the video resolution.
Eg: a 3s video of 360p should lead to output token of 1931*3 output video tokens (based on pricing mentioned here). However, we get back a response containing this:
```
“output_tokens_by_modality”: [
{
“modality”: “video”,
“tokens”: 57920
}
],
```
Seems like it always uses a 10 second video of 720p resolution to calculate the output token, irrespective of what the actual video was? Anyone else seeing a similar issue?
Hello @Shashank_Aggarwal ,
This is fixed now. Thanks for bringing this to our attention!
Today, I requested 360p video, but output has 720p resolution. And charged price is under 720p type.
I feel it was not fixed completely.