Latent Cache Flow: Model-to-Model Communication Without Text

In three linesLatent Cache Flow (LCF) enables direct model-to-model communication via compressed latent representations instead of text. The approach reduces adapter size to 4% of Cache-to-Cache (C2C) and handles differing contexts by transmitting a summary of new information. LCF achieves 23% higher accuracy and 8.5x faster than text-based communication.

Summary generated by Claude — human-verified