Skip to content
Live
QbitAI·· 2 hours ago

Agnes Harness: Dual-system decision-making + dual-layer self-evolution allows Agent efficiency to be increased by up to 14 times, and costs are drastically reduced to 1/27!

Brief overview

Agnes Harness employs Sys1 dual-system decision-making and TTT×RSI double-layer self-evolution mechanism, enabling the 9B decision model Agnes Sys1.0 to collaborate with LLM, improving decision speed by 14.3 times and reducing costs to 1/27 of large models in household leakage tasks.

Source: QbitAI · qbitai.com