arXiv:2606. 19704v1 Announce Type: new Abstract: Agent benchmarks are growing fast, but no single benchmark touches more than four or five of the dimensions that deployment exposes.
arXiv:2602. 03238v3 Announce Type: replace Abstract: LLM agent benchmark scores are shaped not only by the model but also by the agent harness, environment, evaluator, and inference budget.
Agent evaluations increasingly benchmark LLMs, but rankings can be swayed by evaluation conditions such as scaffolds or tasks, making reliability claim‑dependent. A Bayesian variance‑decomposition framework applied to 22 benchmarks shows that reliability varies with the measurement goal: fixed model‑scaffold systems rank reliably, while underlying‑model rankings are less stable. Scaffold choice can alter conclusions, and adding more tasks only modestly improves reliability when scaffold coverage is limited; however, pooling diverse benchmarks can substantially raise cross‑task ranking reliability and reduce cost.
The paper argues that the current LLM-as-a-judge evaluation method, which compensates for systematic measurement bias by increasing the number of comparisons, is statistically unsound and computationally wasteful. It identifies that treating LLM judges as neutral ignores documented biases such as position bias, verbosity bias, judge severity, and self‑enhancement. The authors propose a unified latent variable framework that jointly models pairwise and ordinal data while explicitly correcting for these confounders, enabling reliable rankings with far fewer comparisons and negligible additional compute.
arXiv:2607. 24063v1 Announce Type: new Abstract: On standard factuality tasks, frontier models now cluster near the top of the scale.
arXiv:2608. 11323v1 Announce Type: new Abstract: Enterprise practitioners read agent leaderboards as if they ranked agent capability.
arXiv:2607. 16259v1 Announce Type: new Abstract: Pretrained models are typically ranked on multi-task leaderboards to assess their effectiveness across diverse tasks.
GAUGE is a new offline protocol that evaluates whether the common practice of using an LLM-as-a-judge to rank task‑oriented agents actually aligns with a verifiable reward. Across 25 agents from six providers on two benchmarks, GAUGE finds that user satisfaction scores are largely uncorrelated with task success, and that the judge’s ranking loses precision when agents are closely matched in performance. The study highlights a gap between ranking validity and construct validity in current evaluation practices.
On standard factuality tasks, frontier models now cluster near the top of the scale. The question is therefore shifting from how factual a system is toward how much compute that factuality costs.
arXiv:2605. 17273v3 Announce Type: replace-cross Abstract: State-of-the-Art (SOTA) claims pervade Artificial Intelligence (AI) and Machine Learning (ML) research.
Enterprise practitioners read agent leaderboards as if they ranked agent capability. We show, across three open agent-trace benchmarks (TheAgentCompany, $τ^2$-bench, and AppWorld), that the agent main effect accounts for less than 3% of total variance in every dataset and check type, while the agent-by-task interaction accounts for 7-23%.
AgentJudgeBench is a new benchmark that evaluates the reliability of large language model (LLM) judges on agentic tool‑calling tasks involving workflow directed acyclic graphs (DAGs). It contains 3,808 instances across six DAG topologies and three difficulty tiers, tested with five generators (3B–70B open‑weight models and GPT‑5.4) and six judges (20B to frontier scale) under both paired‑with‑and‑without‑ground‑truth conditions. The study finds that judge alignment degrades with task difficulty, ground‑truth exposure can sometimes hurt alignment, and structured evaluation rubrics provide modest improvements, revealing a structural ceiling that model capacity alone cannot surpass.