Y
Hacker News
new
|
ask
|
show
|
jobs
by
acid__
804 days ago
Top multimodal models have about 200-1000 tokens per image, so the math works out.