Bron
TIL about prompt caching.
If your follow-up prompt is within 5 minutes, it's 90% cheaper.
For input tokens only.
And as long as you don't switch models.
Am I understanding this correctly? https://t.co/kY2KHJZS1z
Van @khemaridh (Khe Hy) op X.
TIL about prompt caching.
If your follow-up prompt is within 5 minutes, it's 90% cheaper.
For input tokens only.
And as long as you don't switch models.
Am I understanding this correctly? https://t.co/kY2KHJZS1z
Van @khemaridh (Khe Hy) op X.
Bron