Daily token consumption on the Venice AI platform surged approximately 25% following increased usage of the DeepSeek model