PinchBench

PinchBench is a dataset/benchmark used to evaluate the performance and efficiency of context management frameworks for long-horizon sessions in large language model (LLM) agents.

Papers using PinchBench (4)

SkillAdaptor: Self-Adapting Skills for LLM Agents from Trajectories2026TokenPilot: Cache-Efficient Context Management for LLM Agents2026KAT-Coder-V2.5 Technical Report2026TwinRouterBench: Fast Static and Live Dynamic Evaluation for Realistic Agentic LLM Routing2026