arxiv arXiv cs.LG · 80d ago · research

TokenPilot: Cache-Efficient Context Management for LLM Agents

from English

TokenPilot reduces inference costs by 61% to 87% in both isolated and continuous modes, outperforming prior systems in cost efficiency while maintaining competitive performance. It uses ingestion-aware compaction and lifecycle-aware eviction to stabilize prompt prefixes and manage context segments efficiently.

Importance 2/3 New feature vs. leaders New harness with differentiators arXiv cs.LG Mistral AI NVIDIA OpenAI AI agents Evaluation & benchmarks Inference efficiency

Read original

arxiv arXiv cs.AI · 80d ago · 2 views

TokenPilot: Cache-Efficient Context Management for LLM Agents

TokenPilot reduces inference costs by 61% to 87% in both isolated and continuous modes, outperforming prior systems in cost efficiency while maintaining competitive performance. It uses ingestion-aware compaction and lifecycle-aware eviction to preserve prompt cache continuity and minimize token footprint without introducing prefix mismatches.

arxiv arXiv cs.CL · 73d ago · 7 views

OpenBioRQ: Benchmark for Agentic Biomedical Research Faithfulness

OpenBioRQ introduces a benchmark of 12,553 unsolved biomedical research questions across 12 domains, designed to test agentic models' faithfulness and abstention. It evaluates models in a tool-using setting without answer keys, using real follow-up evidence rather than parametric knowledge, and reveals significant agentic collapse on the hardest questions where tools are no longer used despite being critical.

arxiv arXiv cs.CL · 73d ago · 1 view

MacAgentBench Launches macOS AI Agent Benchmark

MacAgentBench introduces a comprehensive benchmark with 676 tasks across 25 applications, 60% of which involve both GUI and CLI interactions. It uses deterministic rule-based evaluation and fine-grained multi-checkpoint scoring, revealing that Claude Opus 4.6 on OpenClaw achieves 73.7% Pass@1, primarily due to its skill library rather than framework design.

arxiv arXiv cs.AI · 77d ago · 3 views

Efficient and Sound Probabilistic Verification for AI Agents

A new framework enables secure, probabilistic policy enforcement for AI agents in ambiguous environments. It uses distributionally robust optimization to compute rigorous upper bounds on policy violation probabilities without assuming predicate independence. The method outperforms prior approaches on terminal and tool calling agent benchmarks, improving the security-utility trade-off.

arxiv arXiv cs.LG · 77d ago · 2 views

Marginal Advantage Accumulation for Memory-Driven Agent Self-Evolution

This paper introduces Marginal Advantage Accumulation (MAA), a post-processing architecture that addresses cross-batch inconsistency in memory-driven agent self-evolution. MAA formalizes alignment and comparability as structural conditions, uses differential signals and exponential moving average to accumulate signed evidence per operation, and ensures traceability via semantic identity merging. It outperforms batch-level baselines in 14 out of 16 settings and reduces token consumption by about 75%.