LiveMACE: Process-Aware Evaluation of LLM Agent Capabilities in Evolving Markets
cs.AI, cs.CL
Submitted: 2026-10-07
Updated: 2026-10-07
Code: https://github.com/RaymondVoldemortFu/LiveMACE
License: http://creativecommons.org/licenses/by/4.0/
The gist: Evaluating agents by outcomes alone can obscure the capabilities that produce them.
Abstract
Evaluating agents by outcomes alone can obscure the capabilities that produce them. This problem is especially pronounced in evolving environments, where outcomes reflect a closed-loop interaction between agent behavior and changing external conditions. We introduce LiveMACEBench, a process-aware benchmark that uses live financial markets as a naturally evolving testbed for persistent LLM agents. Five frontier LLMs operate along continuous trajectories under matched Tool Use, Persistent Memory, Rule Following, and Multi-Agent Collaboration configurations. We evaluate them through both realized outcomes and mechanism-specific diagnostics derived from complete decision traces. Across 30 days of live evaluation, we find a pronounced outcome-capability gap: realized returns often diverge from capability-specific measurements, and similar outcomes can arise from markedly different patterns of mechanism use. Trace-level diagnostics further expose distinct bottlenecks across capabilities, demonstrating that mechanism access, effective mechanism use, and downstream performance are not interchangeable measures of agent capability. LiveMACEBench makes this distinction measurable, turning live markets from a performance leaderboard into a diagnostic environment for agent capability
Sources
- AgentBench: Evaluating LLMs as Agents
- WebArena: A Realistic Web Environment for Building Autonomous Agents
- SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
- Gaia2: Benchmarking LLM Agents on Dynamic and Asynchronous Environments
- AI-Trader: Benchmarking Autonomous Agents in Real-Time Financial Markets
- ReAct: Synergizing Reasoning and Acting in Language Models
- DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models
- TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks
- Odysseys: Benchmarking Web Agents on Realistic Long Horizon Tasks
- WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation
- SWE-Marathon: Can Agents Autonomously Complete Ultra-Long-Horizon Software Work?
- FinanceBench: A New Benchmark for Financial Question Answering
- PIXIU: A Large Language Model, Instruction Data and Evaluation Benchmark for Finance
- FinBen: A Holistic Financial Benchmark for Large Language Models
- FinMem: A Performance-Enhanced LLM Trading Agent with Layered Memory and Character Design
- A Multimodal Foundation Agent for Financial Trading: Tool-Augmented, Diversified, and Generalist
- TradingAgents: Multi-Agents LLM Financial Trading Framework
- INVESTORBENCH: A Benchmark for Financial Decision-Making Tasks with LLM-based Agent
- From Knowing to Doing: A Memory-Controlled Benchmark for LLM Trading Agents on Stock Markets
- FinEvolveBench: A Benchmark for Self-Evolving Agents on Low-Repetition Tasks with Implicit Rewards
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems