The paper delivers the first systems characterization of agent memory, with a four-axis taxonomy, phase-aware profiler, evaluation of ten systems on two benchmarks, and ten design recommendations.
Available: https://arxiv.org/abs/2511.07885
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
abstract
Large language model (LLM) queries are predominantly processed by frontier models in centralized cloud infrastructure. Demand growth strains this paradigm faster than providers can scale. Two advances create an opportunity to rethink it: small, local LMs (<=20B active parameters) now achieve competitive performance to frontier models on many tasks, and local accelerators (e.g., Apple M4 Max) can host these models at interactive latencies. This raises the question: can local inference viably redistribute demand from centralized infrastructure? This requires measuring both whether local LMs can accurately answer real-world queries and whether they can do so efficiently on power-constrained devices (e.g., laptops). We propose intelligence per watt (IPW), task accuracy per unit of power, as a unified metric for the capability and efficiency of local inference across model-accelerator configurations. We evaluate 20+ state-of-the-art local LMs, 8 hardware accelerators (local and cloud), and 1M real-world single-turn chat and reasoning queries. For each query, we measure accuracy (local LM win rate against frontier models), energy, latency, and power. We find three key results. First, local LMs successfully answer 88.7% of these queries, with accuracy varying by domain. Second, longitudinal analysis from 2023-2025 shows IPW improved 5.3x, driven by both algorithmic and accelerator advances, with locally-serviceable query coverage rising from 23.2% to 71.3%. Third, local accelerators achieve at least 1.4x lower IPW than cloud accelerators running identical models, revealing significant headroom for local accelerator optimization. These findings demonstrate that local inference can meaningfully redistribute demand from centralized infrastructure for a substantial subset of queries, with IPW serving as the critical metric for tracking this transition.
years
2026 7representative citing papers
Introduces nexbax, a diagnostic framework with three themes and 10 dimensions for evaluating AI economic viability, operational practicality, and societal integrity in next-billion-user contexts.
LlamaWeb is a WebGPU backend for llama.cpp that uses static memory planning, tunable kernels, and templated multi-precision support to cut memory use by 29-33% and raise decode throughput by 45-69% versus prior browser frameworks on tested hardware.
MoE structure, not parameter count, governs INT4 robustness in sub-3B VLMs; SigLIP INT8 latency spikes on Jetson Ampere are a BitsAndBytes-Ampere interaction, and INT4 VRAM savings come with TPOT and energy penalties.
Quantitative benchmarks across recent AI accelerators reveal that optimal hardware choice varies with workload parameters and that several platforms incur substantially higher idle power than GPUs.
BaseRT achieves up to 1.56x higher LLM decode throughput than llama.cpp on Apple Silicon through native Metal kernel fusion and unified memory optimizations.
AgentStop uses execution signals to early-terminate failing local LLM agent trajectories, cutting energy use 15-20% with minimal utility loss.
citing papers explorer
-
Agent Memory: Characterization and System Implications of Stateful Long-Horizon Workloads
The paper delivers the first systems characterization of agent memory, with a four-axis taxonomy, phase-aware profiler, evaluation of ten systems on two benchmarks, and ten design recommendations.
-
Next-Billion AI Index: The compass for AI utility and adoption in the global majority
Introduces nexbax, a diagnostic framework with three themes and 10 dimensions for evaluating AI economic viability, operational practicality, and societal integrity in next-billion-user contexts.
-
Llamas on the Web: Memory-Efficient, Performance-Portable, and Multi-Precision LLM Inference with WebGPU
LlamaWeb is a WebGPU backend for llama.cpp that uses static memory planning, tunable kernels, and templated multi-precision support to cut memory use by 29-33% and raise decode throughput by 45-69% versus prior browser frameworks on tested hardware.
-
Rethinking Small VLM Quantization: From Component-Wise Analysis to Hardware-Aware Edge Deployment
MoE structure, not parameter count, governs INT4 robustness in sub-3B VLMs; SigLIP INT8 latency spikes on Jetson Ampere are a BitsAndBytes-Ampere interaction, and INT4 VRAM savings come with TPOT and energy penalties.
-
The xPU-athalon: Quantifying the Competition of AI Acceleration
Quantitative benchmarks across recent AI accelerators reveal that optimal hardware choice varies with workload parameters and that several platforms incur substantially higher idle power than GPUs.
-
BaseRT: Best-in-Class LLM Inference on Apple Silicon via Native Metal
BaseRT achieves up to 1.56x higher LLM decode throughput than llama.cpp on Apple Silicon through native Metal kernel fusion and unified memory optimizations.
-
AgentStop: Terminating Local AI Agents Early to Save Energy in Consumer Devices
AgentStop uses execution signals to early-terminate failing local LLM agent trajectories, cutting energy use 15-20% with minimal utility loss.