Agent Benchmarks

Fine-grained recognition in everyday life is often not a closed-book classification problem: when encountering unfamiliar objects, humans actively search, compare visual details, and verify evidence before deciding. Existing benchmarks primarily evaluate visually recognition, leaving this active external knowledge acquisition ability underexplored. We study fine-grained knowledge acquisition, where a system must seek, verify, and use external evidence to answer open-ended fine-grained recognition questions. We introduce FIKA-Bench, a leakage-aware and evidence-grounded collection of 311 public-source and real-life instances. To ensure high quality, every example is filtered against frontier closed-book models to remove memorized cases and audited to eliminate image-answer leakage, retaining only samples supported by verified evidence. Our evaluation of latest Large Multimodal Models (LMMs) and agents reveals that the task remains a formidable challenge: the best system reaches only 25.1% accuracy, with no model exceeding 30%. Crucially, we find that merely equipping models with tools is insufficient to bridge this gap; agent failures are predominantly driven by wrong entity retrieval and poor visual judgement. These results show that reliable knowledge acquisition needs better agent designs that focus on fine-grained recognition.

Long-term memory is crucial for agents in specialized web environments, where success depends on recalling interface affordances, state dynamics, workflows, and recurring failure modes. However, existing memory benchmarks for agents mostly focus on user histories, short traces, or downstream task success, leaving open how to directly evaluate whether memory systems effectively internalize environment-specific experience. To address this gap, we introduce LongMemEval-V2 (LME-V2), a benchmark for evaluating whether memory systems can help agents acquire the experience needed to become knowledgeable colleagues in customized environments. LME-V2 contains 451 manually curated questions covering five core memory abilities for web agents: static state recall, dynamic state tracking, workflow knowledge, environment gotchas, and premise awareness. Questions are paired with history trajectories containing up to 500 trajectories and 115M tokens. We use a context gathering formulation: memory systems consume history trajectories and return compact evidence for downstream question answering. We propose a suite of two memory methods: AgentRunbook-R, an efficient RAG-based memory with knowledge pools for raw state observations, events, and strategy notes, and AgentRunbook-C, which stores trajectories as files and invokes a coding agent to gather evidence in an augmented sandbox. Experiments show that AgentRunbook-C achieves the best performance with 72.5% average accuracy, outperforming the strongest RAG baseline (48.5%) and the off-the-shelf coding agent baseline (69.3%). Despite the strong performance gains, coding agent based methods have high latency costs. While AgentRunbook-C advances the accuracy-latency Pareto frontier, substantial room for improvement remains. Together, these results establish LME-V2 as a challenging testbed for developing long-term memory systems for environment experience.

The large-scale deployment of personalized healthcare agents demands memory mechanisms that are exceptionally precise, safe, and capable of long-term clinical tracking. However, existing benchmarks primarily focus on daily open-domain conversations, failing to capture the high-stakes complexity of real-world medical applications. Motivated by the stringent production requirements of an industry-leading health management agent serving tens of millions of active users, we introduce MedMemoryBench. We develop a human-agent collaborative pipeline to synthesize highly realistic, long-horizon medical trajectories based on clinically grounded, synthetic patient archetypes. This process yields a massive, expertly validated dataset comprising approximately 2,000 sessions and 16,000 interaction turns. Crucially, MedMemoryBench departs from traditional static evaluations by pioneering an "evaluate-while-constructing" streaming assessment protocol, which precisely mirrors dynamic memory accumulation in production environments. Furthermore, we formalize and systematically investigate the critical phenomenon of memory saturation, where sustained information influx actively degrades retrieval and reasoning robustness. Comprehensive benchmarking reveals severe bottlenecks in mainstream architectures, particularly concerning complex medical reasoning and noise resilience. By exposing these fundamental flaws, MedMemoryBench establishes a vital foundation for developing robust, production-ready medical agents.

Clinical reasoning agents based on large language models (LLMs) aim to automate tasks such as intensive care unit (ICU) monitoring and patient state tracking from electronic health records (EHRs). Existing systems typically rely on manually curated clinical tools or skills for concepts such as sepsis detection and organ failure assessment. However, maintaining these tool libraries requires substantial expert effort, while zero-shot querying or code generation often produces inefficient and unreliable reasoning chains, especially under institution-specific clinical policies. We introduce CodeClinic, a benchmark built on MIMIC-IV for evaluating whether LLM agents can synthesize and compose reusable clinical skills instead of relying on fixed toolboxes. The benchmark contains two complementary tasks: longitudinal ICU surveillance and compositional information seeking. The longitudinal setting simulates monitoring patient trajectories with structured decisions every four hours across 25 findings and eight clinical families, while the compositional setting spans 63k instances across 259 tasks in nine domains and is stratified by compositional dependency depth to evaluate increasingly complex multi-step reasoning. We further propose an offline autoformalization pipeline that converts natural-language clinical guidelines into reusable and verified Python skill libraries through iterative LLM refinement. Compared with zero-shot code generation, the resulting libraries improve consistency while reducing per-query token usage by up to 40%.

Most LLM benchmarks score how well a model responds to explicit requests. They leave unmeasured a different conversational ability: noticing and acting on needs the user has implied but not said. We call this conversational proactivity. ProactBench decomposes it into three phase-tied types: Emergent, inference from a single disclosed anchor; Critical, synthesis across multiple anchors; and Recovery, grounded forward-looking value after task completion. We operationalise the benchmark with three agents: a Planner, a User Agent, and an Assistant Model. Their information asymmetries defend against style-confounded scoring, rubric leakage, external-context contamination, and information dumps. The released corpus contains 198 curated dialogues with 624 trigger points across 24 communication styles drawn from a psychometric inventory and audited by an independent LLM judge. Across 16 frontier and open-weight models, Recovery is both difficult and weakly predicted by six standard benchmarks, making it a useful new evaluation signal.

Autonomous data analysis agents are increasingly expected to conduct exploratory analysis with limited human guidance about data. However, existing benchmarks typically evaluate such agents in prior-guided settings, providing selected data sources, explicit data schemas, or cleaned data, thereby understating the exploratory burden. To evaluate this realistic exploratory data analysis task, we introduce DataClawBench, a benchmark built from financial think-tank consulting scenarios where agents must independently explore unfamiliar, noisy, cross-domain data and produce verifiable conclusions. DataClawBench provides a unified real-world data environment with approximately 2.06 million records across enterprise, industry, and policy domains, with native data noise preserved. On top of this data environment, it defines 492 multi-step cross-domain tasks, each annotated with intermediate milestones that diagnose exploration and reasoning failures beyond outcome accuracy. A systematic evaluation of eight advanced LLMs under the OpenClaw agent reveals that exploratory data analysis breaks agent reliability: more exploration does not reliably translate into task-relevant progress or correct final answers.

The rapid advancement of Multimodal Large Language Models (MLLMs) has empowered Unmanned Aerial Vehicle (UAV) with exceptional capabilities in spatial reasoning, semantic understanding, and complex decision-making, making them inherently suited for UAV Search and Rescue (SAR). However, existing UAV SAR research is dominated by traditional vision and path-planning methods and lacks a comprehensive and unified benchmark for embodied agents. To bridge this gap, we first propose the novel task of Embodied Search and Rescue (ESAR), which requires aerial agents to autonomously explore complex environments, identify rescue clues, and reason about victim locations to execute informed decision-making. Additionally, we present ESARBench, the first comprehensive benchmark designed to evaluate MLLM-driven UAV agents in highly realistic SAR scenarios. Leveraging Unreal Engine 5 and AirSim, we construct four high-fidelity, large-scale open environments mapped directly from real-world Geographic Information System (GIS) data to ensure photorealistic landscapes. To rigorously simulate actual rescue operations, our benchmark incorporates dynamic variables including weather conditions, time of day, and stochastic clue placement. Furthermore, we create a dataset of 600 tasks modeled after real-world rescue cases and propose a robust set of evaluation metrics. We evaluate diverse baselines, ranging from traditional heuristics to advanced ground and aerial MLLM-based ObjectNav agents. Experimental results highlight the challenges in ESAR, revealing critical bottlenecks in spatial memory, aerial adaptation, and the trade-off between search efficiency and flight safety. We hope ESARBench serves as a valuable resource to advance research on Embodied Search and Rescue domain. Source code and project page: https://4amgodvzx.github.io/ESAR.github.io.

Earth Observation (EO) analysis is inherently interactive: resolving uncertainty often requires expanding the region of interest, retrieving historical observations, and switching across sensors such as optical and Synthetic Aperture Radar. However, most EO benchmarks collapse this process into fixed-input, single-turn tasks. To address this gap, we present EO-Gym, a controlled executable framework for multimodal, tool-using EO agents that formulates EO analysis as a Gymnasium-style local geospatial workspace backed by more than 660k multimodal files indexed by location, time, and sensor type, with 35 EO-specialized tools spanning six task families. Built on this environment, we construct EO-Gym-Data, a benchmark of 9,078 trajectories and 34,604 reasoning steps, and grounded in eight public EO datasets together with Landsat and Sentinel-2 imagery. Evaluating $10$ open and closed VLMs shows that strong general-purpose models still struggle with interactive EO reasoning, especially on temporal and cross-modal workflows. As a reference baseline, EO-Gym-4B, obtained by fine-tuning Qwen3-VL-4B-Instruct on EO-Gym-Data, improves overall Pass@3 from $0.49$ to $0.74$ under the main evaluation setting. O-Gym provides a reproducible environment for interactive EO agents, operationalizing EO as an evidence-gathering problem that requires planning across geospatial, temporal, and sensing modality.

Clinical reasoning demands multi-step interactions -- gathering patient history, ordering tests, interpreting results, and making safe treatment decisions -- yet a unified training environment provides the breadth of clinical domains and specialized tools to train generalizable medical AI agents through reinforcement learning remains elusive. We present a comprehensive empirical study of multi-turn agentic RL for medical AI, built on \gym{}, a gymnasium-compatible environment spanning 10 clinical domains with 3.6K+ tasks, 135 domain-specific tools, and a knowledge base of 828K medical passages. Our analysis reveals that agentic multi-turn structure degrades into verbose single-turn monologues, characterized by monotonic length explosion and a simultaneous erosion of tool-use frequency. We characterize how this collapse, alongside distillation instability, stems from the misalignment of sparse terminal rewards with sequential clinical trajectories. We find that vanilla GRPO achieves strong final accuracy on some benchmarks but suffers from training instability, evidenced by significant oscillations in response length and prolonged convergence periods. To improve training efficiency and stability, we propose Turn-level Truncated On-Policy Distillation (TT-OPD), a self-distillation framework where a gradient-free EMA teacher leverages outcome-privileged information to provide dense, outcome-aware KL regularization at every conversation turn. TT-OPD achieves the best performance on 10 of 18 benchmarks with an average +3.9~pp improvement over the non-RL baseline with faster early convergence, controlled response length, and sustained multi-turn tool use.

We introduce LATTICE, a benchmark for evaluating the decision support utility of crypto agents in realistic user-facing scenarios. Prior crypto agent benchmarks mainly focus on reasoning-based or outcome-based evaluation, but do not assess agents' ability to assist user decision-making. LATTICE addresses this gap by: (1) defining six evaluation dimensions that capture key decision support properties; (2) proposing 16 task types that span the end-to-end crypto copilot workflow; and (3) using LLM judges to automatically score agent outputs based on these dimensions and tasks. Crucially, the dimensions and tasks are designed to be evaluable at scale using LLM judges, without relying on ground truth from expert annotators or external data sources. In lieu of these dependencies, LATTICE's LLM judge rubrics can be continually audited and updated given new dimensions, tasks, criteria, and human feedback, thus promoting reliable and extensible evaluation. While other benchmarks often compare foundation models sharing a generic agent framework, we use LATTICE to assess production-level agents used in actual crypto copilot products, reflecting the importance of orchestration and UI/UX design in determining agent quality. In this paper, we evaluate six real-world crypto copilots on 1,200 diverse queries and report breakdowns across dimensions, tasks, and query categories. Our experiments show that most of the tested copilots achieve comparable aggregate scores, but differ more significantly on dimension-level and task-level performance. This pattern suggests meaningful trade-offs in decision support quality: users with different priorities may be better served by different copilots than the aggregate rankings alone would indicate. To support reproducible research, we open-source all LATTICE code and data used in this paper.

Real-world data visualization (DV) requires native environmental grounding, cross-platform evolution, and proactive intent alignment. Yet, existing benchmarks often suffer from code-sandbox confinement, single-language creation-only tasks, and assumption of perfect intent. To bridge these gaps, we introduce DV-World, a benchmark of 260 tasks designed to evaluate DV agents across real-world professional lifecycles. DV-World spans three domains: DV-Sheet for native spreadsheet manipulation including chart and dashboard creation as well as diagnostic repair; DV-Evolution for adapting and restructuring reference visual artifacts to fit new data across diverse programming paradigms and DV-Interact for proactive intent alignment with a user simulator that mimics real-world ambiguous requirements. Our hybrid evaluation framework integrates Table-value Alignment for numerical precision and MLLM-as-a-Judge with rubrics for semantic-visual assessment. Experiments reveal that state-of-the-art models achieve less than 50% overall performance, exposing critical deficits in handling the complex challenges of real-world data visualization. DV-World provides a realistic testbed to steer development toward the versatile expertise required in enterprise workflows. Our data and code are available at this project page.

Agentic AI systems are rapidly advancing toward real-world applications, yet their readiness in complex and personalized environments remains insufficiently characterized. To address this gap, we introduce PersonalHomeBench, a benchmark for evaluating foundation models as agentic assistants in personalized smart home environments. The benchmark is constructed through an iterative process that progressively builds rich household states, which are then used to generate personalized, context-dependent tasks. To support realistic agent-environment interaction, we provide PersonalHomeTools, a comprehensive toolbox enabling household information retrieval, appliance control, and situational understanding. PersonalHomeBench evaluates both reactive and proactive agentic abilities under unimodal and multimodal observations. Thorough experimentation reveals a systematic performance reduction as task complexity increases, with pronounced failures in counterfactual reasoning and under partial observability, where effective tool-based information gathering is required. These results position PersonalHomeBench as a rigorous evaluation platform for analyzing the robustness and limitations of personalized agentic reasoning and planning.

The development of general-purpose agents requires a shift from executing simple instructions to completing complex, real-world productivity workflows. However, current tool-use benchmarks remain misaligned with real-world requirements, relying on AI-generated queries, dummy tools, and limited system-level coordination. To address this, we propose GTA-2, a hierarchical benchmark for General Tool Agents (GTA) spanning atomic tool use and open-ended workflows. Built on real-world authenticity, it leverages real user queries, deployed tools, and multimodal contexts. (i) GTA-Atomic, inherited from our prior GTA benchmark, evaluates short-horizon, closed-ended tool-use precision. (ii) GTA-Workflow introduces long-horizon, open-ended tasks for realistic end-to-end completion. To evaluate open-ended deliverables, we propose a recursive checkpoint-based evaluation mechanism that decomposes objectives into verifiable sub-goals, enabling unified evaluation of both model capabilities and agent execution frameworks (i.e., execution harnesses). Experiments reveal a pronounced capability cliff: while frontier models already struggle on atomic tasks (below 50%), they largely fail on workflows, with top models achieving only 14.39% success. Further analysis shows that checkpoint-guided feedback improves performance, while advanced frameworks such as Manus and OpenClaw substantially enhance workflow completion, highlighting the importance of execution harness design beyond the underlying model capacity. These findings provide guidance for developing reliable personal and professional assistants. Dataset and code will be available at https://github.com/open-compass/GTA.

As agentic foundation models continue to evolve, how to further improve their performance in vertical domains has become an important challenge. To this end, building upon Tongyi DeepResearch, a powerful agentic foundation model, we focus on the Chinese medical deep search scenario and propose QuarkMedSearch, systematically exploring a full-pipeline approach spanning medical multi-hop data construction, training strategies, and evaluation benchmarks to further push and assess its performance upper bound in vertical domains. Specifically, for data synthesis, to address the scarcity of deep search training data in the medical domain, we combine a large-scale medical knowledge graph with real-time online exploration to construct long-horizon medical deep search training data; for post-training, we adopt a two-stage SFT and RL training strategy that progressively enhances the model's planning, tool invocation, and reflection capabilities required for deep search, while maintaining search efficiency; for evaluation, we collaborate with medical experts to construct the QuarkMedSearch Benchmark through rigorous manual verification. Experimental results demonstrate that QuarkMedSearch achieves state-of-the-art performance among open-source models of comparable scale on the QuarkMedSearch Benchmark, while also maintaining strong competitiveness on general benchmarks.

Existing person search methods assume access to complete visual queries or exhaustive tracking, yet real-world witness accounts are vague, partial, and spread across cameras and time. We introduce ARGOS (Agentic Retrieval with Grounded Observational Search), a benchmark and agent framework that recasts multi-camera person search from one-shot retrieval on a complete query into interactive reasoning from partial clues. To our knowledge, ARGOS is the first interactive benchmark to couple witness dialogue with camera-network topology, requiring an agent to plan, question, and eliminate under information asymmetry. An ARGOS agent receives a vague witness statement and must decide what to ask, when to invoke spatial or temporal tools, and how to interpret ambiguous natural-language responses, all within a limited turn budget. To ground reasoning in physical constraints, the agent accesses a Spatio-Temporal Topology Graph (STTG) encoding camera connectivity and empirically validated transition times. The benchmark comprises 2{,}691 tasks across 14 real-world scenarios in three progressive tracks: semantic perception (Who, 989 tasks), spatial reasoning (Where, 550 tasks), and temporal reasoning (When, 1{,}152 tasks). We propose Turn-Weighted Success (TWS) as the primary metric, jointly measuring correctness and turn efficiency. Experiments with four LLM backbones show the benchmark is far from solved: the best agent achieves TWS of 0.383 (Track2) and 0.590 (Track3). Ablations confirm each component is essential: removing domain-specific tools drops Top-1 accuracy by up to 49.6 percentage points, and removing strategic reasoning halves TWS while barely affecting Top-1.

We describe compute-grounded reasoning (CGR), a design pattern in which code computes selected sub-problems from explicit intermediate representations before a language model answers. Spatial Atlas implements CGR as an Agent2Agent (A2A) server with a spatial question-answering handler and a machine-learning engineering handler. The spatial handler asks a language model to extract a scene graph, and code then fills in missing distances and checks the extracted safety rules. A separate benchmark driver can also run a strict metric bridge. It computes the gap for horizontal-gap questions from segmentation masks and a reconstructed point map, and it passes that gap to the answering model as a fact. The bridge returns a fixed unavailable answer when an evidence check fails, and it never falls back to model-estimated coordinates. The ML-engineering handler generates pipeline code, parses validation scores, and caps the number of repair and refinement passes. Its code execution is off by default. The repository also provides four run modes that can write label-free journals, a shuffled-image control mapping, and journal validators that reject label-bearing fields. We report one private label-free operational run in which four paths each wrote eight prediction rows with zero retries. Labels stayed sealed, and no score was computed, so this run establishes operational integrity only. We report no FieldWorkArena result because the benchmark data were not accessible. We also omit every performance, latency, and resource-use number that lacks a reproducible run artifact.

ATANT v1.0 (arXiv:2604.06710) defined continuity as a system property with 7 required properties and introduced a 10-checkpoint, LLM-free evaluation methodology validated on a 250-story corpus. Since publication, a recurring reviewer and practitioner question has concerned not the framework itself but its relationship to a wider set of memory evaluations: LOCOMO, LongMemEval, BEAM, MemoryBench, Zep's evaluation suite, Letta/MemGPT's evaluations, and RULER. This companion paper, v1.1, does not modify the v1.0 standard. It closes a related-work gap that v1.0 left brief under page limits. We show by structural analysis that none of these benchmarks measures continuity as defined in v1.0: of the 7 required properties, the median existing eval covers 1 property, the mean covers 0.43 when partial credit is scored at 0.5, and no eval covers more than 2. We provide a cell-by-cell property-coverage matrix, identify methodological defects specific to each benchmark (including an empty-gold scoring bug in the LOCOMO reference implementation that renders 23% of its corpus unscorable by construction), and publish our reference implementation's LOCOMO score (8.8%) alongside the structural reason that number is uninformative about continuity. We publish our 8.8% LOCOMO score alongside our 96% ATANT cumulative-scale score as a calibration pair: the 87-point divergence is evidence that the two benchmarks measure different properties, not that one system is an order of magnitude better than another. The position v1.1 takes is not adversarial: each benchmark measures a real capability. The claim is that none of them can adjudicate continuity, and conflating them with continuity evaluation has led the field to under-invest in the properties v1.0 names.

Proactivity is a core expectation for AGI. Prior work remains largely confined to laboratory settings, leaving a clear gap in real-world proactive agent: depth, complexity, ambiguity, precision and real-time constraints. We study this setting, where useful intervention requires inferring latent needs from ongoing context and grounding actions in evolving user memory under latency and long-horizon constraints. We first propose DD-MM-PAS (Demand Detection, Memory Modeling, Proactive Agent System) as a general paradigm for streaming proactive AI agent. We instantiate this paradigm in Pask, with streaming IntentFlow model for DD, a hybrid memory (workspace, user, global) for long-term MM, PAS infra framework and introduce how these components form a closed loop. We also introduce LatentNeeds-Bench, a real-world benchmark built from user-consented data and refined through thousands of rounds of human editing. Experiments show that IntentFlow matches leading Gemini3-Flash models under latency constraints, while identifying deeper user intent.

Existing Agent benchmarks suffer from two critical limitations: high environment interaction overhead (up to 41% of total evaluation time) and imbalanced task horizon and difficulty distributions that make aggregate scores unreliable. To address these issues, we propose AgentCE-Bench built around a unified grid-based planning task, where agents must fill hidden slots in a partially completed schedule subject to both local slot constraints and global constraints. Our benchmark offers fine-grained control through two orthogonal axes: Scalable Horizons, controlled by the number of hidden slots $H$, and Controllable Difficulty, governed by a decoy budget $B$ that determines the number of globally misleading decoy candidates. Crucially, all tool calls are resolved via static JSON files under a Lightweight Environment design, eliminating setup overhead and enabling fast, reproducible evaluation suitable for training-time validation. We first validate that $H$ and $B$ provide reliable control over task horizon and difficulty, and that AgentCE-Bench exhibits strong domain consistency and model discriminability. We then conduct comprehensive experiments across 13 models of diverse sizes and families over 6 domains, revealing significant cross-model performance variation and confirming that AgentCE-Bench provides interpretable and controllable evaluation of agent reasoning.

As the focus in LLM-based coding shifts from static single-step code generation to multi-step agentic interaction with tools and environments, understanding which tasks will challenge agents and why becomes increasingly difficult. This is compounded by current practice: agent performance is typically measured by aggregate pass rates on benchmarks, but single-number metrics obscure the diversity of tasks within a benchmark. We present a framework for predicting success or failure on individual tasks tailored to the agentic coding regime. Our approach augments Item Response Theory (IRT) with rich features extracted from tasks, including issue statements, repository contexts, solutions, and test cases, and introduces a novel decomposition of agent ability into LLM and scaffold ability components. This parameterization enables us to aggregate evaluation data across heterogeneous leaderboards and accurately predict task-level performance for unseen benchmarks, as well as unseen LLM-scaffold combinations. Our methods have practical utility for benchmark designers, who can better calibrate the difficulty of their new tasks without running computationally expensive agent evaluations.

Agentic vision-language models increasingly act through extended interactions, but most evaluations still focus on single-image, single-turn correctness. We introduce AMIGO (Agentic Multi-Image Grounding Oracle Benchmark), a long-horizon benchmark for hidden-target identification over galleries of visually similar images. In AMIGO, the oracle privately selects a target image, and the model must recover it by asking a sequence of attribute-focused Yes/No questions under a strict protocol that returns Yes/No/Unsure feedback and penalizes invalid actions with Skip. This setting stresses (i) question selection under uncertainty, (ii) consistent constraint tracking across turns, and (iii) fine-grained discrimination as evidence accumulates. We instantiate AMIGO with the Guess My Preferred Dress task and evaluate open-source VLMs with metrics covering identification success, evidence verification, efficiency, protocol compliance, robustness to controlled feedback perturbations, and trajectory-level diagnostics. The benchmarking results show that final-answer accuracy alone overstates evidence-grounded performance: models can guess correctly without verification-passing evidence, waste turns through invalid questions, or fail to preserve the upload protocol. Strong AMIGO performance depends on the combination of visual discrimination, informative question selection, constraint tracking, efficient stopping, sustained protocol following, and recovery from controlled feedback noise; model scale alone does not guarantee reliable long-horizon interactive grounding.

We introduce ARC-AGI-3, an interactive benchmark for studying agentic intelligence through novel, abstract, turn-based environments in which agents must explore, infer goals, build internal models of environment dynamics, and plan effective action sequences without explicit instructions. Like its predecessors ARC-AGI-1 and 2, ARC-AGI-3 focuses entirely on evaluating fluid adaptive efficiency on novel tasks, while avoiding language and external knowledge. ARC-AGI-3 environments only leverage Core Knowledge priors and are difficulty-calibrated via extensive testing with human test-takers. Our testing shows humans can solve 100% of the environments, in contrast to frontier AI systems which, as of March 2026, score below 1%. In this paper, we present the benchmark design, its efficiency-based scoring framework grounded in human action baselines, and the methodology used to construct, validate, and calibrate the environments.

We introduce GTO Wizard Benchmark, a public API and standardized evaluation framework for benchmarking algorithms in Heads-Up No-Limit Texas Hold'em (HUNL). The benchmark evaluates agents against GTO Wizard AI, a state-of-the-art superhuman poker agent that approximates Nash Equilibria, and defeated Slumbot, the 2018 Annual Computer Poker Competition champion and previous strongest publicly accessible HUNL benchmark, by $19.4$ $\pm$ $4.1$ bb/100. Variance is a fundamental challenge in poker evaluation; we address this by integrating AIVAT, a provably unbiased variance reduction technique that achieves equivalent statistical significance with ten times fewer hands than naive Monte Carlo evaluation. We conduct a comprehensive benchmarking study of state-of-the-art large language models under zero-shot conditions, including GPT-5.4, Claude Opus 4.6, Gemini 3.1 Pro, Grok 4, and others. Initial results and analysis reveal dramatic progress in LLM reasoning over recent years, yet all models remain far below the baseline established by our benchmark. Qualitative analysis reveals clear opportunities for improvement, including representation and the ability to reason over hidden states. This benchmark provides researchers with a precise and quantifiable setting to evaluate advances in planning and reasoning in multi-agent systems with partial observability.

Multimodal AI agents are increasingly automating complex real-world workflows that involve online web execution. However, current web-agent benchmarks suffer from a critical limitation: they focus entirely on web-based interaction and perception, lacking grounding in the user's real-world physical surroundings. This limitation prevents evaluation in crucial scenarios, such as when an agent must use egocentric visual perception (e.g., via AR glasses) to recognize an object in the user's surroundings and then complete a related task online. To address this gap, we introduce Ego2Web, the first benchmark designed to bridge egocentric video perception and web agent execution. Ego2Web pairs real-world first-person video recordings with web tasks that require visual understanding, web task planning, and interaction in an online environment for successful completion. We utilize an automatic data-generation pipeline combined with human verification and refinement to curate well-constructed, high-quality video-task pairs across diverse web task types, including e-commerce, media retrieval, knowledge lookup, etc. To facilitate accurate and scalable evaluation for our benchmark, we also develop a novel LLM-as-a-Judge automatic evaluation method, Ego2WebJudge, which achieves approximately 84% agreement with human judgment, substantially higher than existing evaluation methods. Experiments with diverse SoTA agents on our Ego2Web show that their performance is weak, with substantial headroom across all task categories. We also conduct a comprehensive ablation study on task design, highlighting the necessity of accurate video understanding in the proposed task and the limitations of current agents. We hope Ego2Web can be a critical new resource for developing truly capable AI assistants that can seamlessly see, understand, and act across the physical and digital worlds.

While Large Language Models (LLMs) have evolved into tool-using agents, they remain brittle in long-horizon interactions. Unlike mathematical reasoning where errors are often rectifiable via backtracking, tool-use failures frequently induce irreversible side effects, making accurate step-level verification critical. However, existing process-level benchmarks are predominantly confined to closed-world mathematical domains, failing to capture the dynamic and open-ended nature of tool execution. To bridge this gap, we introduce AgentProcessBench, the first benchmark dedicated to evaluating step-level effectiveness in realistic, tool-augmented trajectories. The benchmark comprises 1,000 diverse trajectories and 8,509 human-labeled step annotations with 89.1% inter-annotator agreement. It features a ternary labeling scheme to capture exploration and an error propagation rule to reduce labeling ambiguity. Extensive experiments reveal key insights: (1) weaker policy models exhibit inflated ratios of correct steps due to early termination; (2) distinguishing neutral and erroneous actions remains a significant challenge for current models; and (3) process-derived signals provide complementary value to outcome supervision, significantly enhancing test-time scaling. We hope AgentProcessBench can foster future research in reward models and pave the way toward general agents. The code and data are available at https://github.com/RUCBM/AgentProcessBench.

Autonomous coding agents can produce strong tabular baselines quickly on Kaggle-style tasks. Practical value depends on end-to-end correctness and reliability under time limits. This paper introduces TML-Bench, a tabular benchmark for data science agents on Kaggle-style tasks. This paper evaluates 10 OSS LLMs on four Kaggle competitions and three time budgets (240s, 600s, and 1200s). Each model is run five times per task and budget. A run is successful if it produces a valid submission and a private-holdout score on hidden labels that are not accessible to the agent. This paper reports median performance, success rates, and run-to-run variability. MiniMax-M2.1 model achieves the best aggregate performance score on all four competitions under the paper's primary aggregation. Average performance improves with larger time budgets. Scaling is noisy for some individual models at the current run count. Code and materials are available at https://github.com/MykolaPinchuk/TML-bench/tree/master.

We present a system for training enterprise search agents via reinforcement learning that achieves state-of-the-art performance across a diverse suite of hard-to-verify agentic search tasks. Our work makes four core contributions. First, we introduce KARLBench, a multi-capability evaluation suite spanning six distinct search regimes, including constraint-driven entity search, cross-document report synthesis, tabular numerical reasoning, exhaustive entity retrieval, procedural reasoning over technical documentation, and fact aggregation over internal enterprise notes. Second, we show that models trained across heterogeneous search behaviors generalize substantially better than those optimized for any single benchmark. Third, we develop an agentic synthesis pipeline that employs long-horizon reasoning and tool use to generate diverse, grounded, and high-quality training data, with iterative bootstrapping from increasingly capable models. Fourth, we propose a new post-training paradigm based on iterative large-batch off-policy RL that is sample efficient, robust to train-inference engine discrepancies, and naturally extends to multi-task training with out-of-distribution generalization. Compared to Claude 4.6 and GPT 5.2, KARL is Pareto-optimal on KARLBench across cost-quality and latency-quality trade-offs, including tasks that were out-of-distribution during training. With sufficient test-time compute, it surpasses the strongest closed models. These results show that tailored synthetic data in combination with multi-task reinforcement learning enables cost-efficient and high-performing knowledge agents for grounded reasoning.

LLM agents are rapidly becoming the practical interface for task automation, yet the ecosystem lacks a principled way to choose among an exploding space of deployable configurations. Existing LLM leaderboards and tool/agent benchmarks evaluate components in isolation and remain fragmented across tasks, metrics, and candidate pools, leaving a critical research gap: there is little query-conditioned supervision for learning to recommend end-to-end agent configurations that couple a backbone model with a toolkit. We address this gap with AgentSelect, a benchmark that reframes agent selection as narrative query-to-agent recommendation over capability profiles and systematically converts heterogeneous evaluation artifacts into unified, positive-only interaction data. AgentSelectcomprises 111,179 queries, 107,721 deployable agents, and 251,103 interaction records aggregated from 40+ sources, spanning LLM-only, toolkit-only, and compositional agents. Our analyses reveal a regime shift from dense head reuse to long-tail, near one-off supervision, where popularity-based CF/GNN methods become fragile and content-aware capability matching is essential. We further show that Part~III synthesized compositional interactions are learnable, induce capability-sensitive behavior under controlled counterfactual edits, and improve coverage over realistic compositions; models trained on AgentSelect also transfer to a public agent marketplace (MuleRun), yielding consistent gains on an unseen catalog. Overall, AgentSelect provides the first unified data and evaluation infrastructure for agent recommendation, which establishes a reproducible foundation to study and accelerate the emerging agent ecosystem.

As autonomous systems such as drones, become increasingly deployed in high-stakes, human-centric domains, it is critical to evaluate the ethical alignment since failure to do so imposes imminent danger to human lives, and long term bias in decision-making. Automated ethical benchmarking of these systems is understudied due to the lack of ubiquitous, well-defined metrics for evaluation, and stakeholder-specific subjectivity, which cannot be modeled analytically. To address these challenges, we propose SEED-SET, a Bayesian experimental design framework that incorporates domain-specific objective evaluations, and subjective value judgments from stakeholders. SEED-SET models both evaluation types separately with hierarchical Gaussian Processes, and uses a novel acquisition strategy to propose interesting test candidates based on learnt qualitative preferences and objectives that align with the stakeholder preferences. We validate our approach for ethical benchmarking of autonomous agents on two applications and find our method to perform the best. Our method provides an interpretable and efficient trade-off between exploration and exploitation, by generating up to $2\times$ optimal test candidates compared to baselines, with $1.25\times$ improvement in coverage of high dimensional search spaces.

The research community lacks a middle ground between StarCraft II full game and its mini-games. The full-game's sprawling state-action space renders reward signals sparse and noisy, but in mini-games simple agents saturate performance. This complexity gap hinders steady curriculum design and prevents researchers from experimenting with modern Reinforcement Learning algorithms in RTS environments under realistic compute budgets. To fill this gap, we present the Two-Bridge Map Suite, the first entry in an open-source benchmark series we purposely engineered as an intermediate benchmark to sit between these extremes. By disabling economy mechanics such as resource collection, base building, and fog-of-war, the environment isolates two core tactical skills: long-range navigation and micro-combat. Preliminary experiments show that agents learn coherent maneuvering and engagement behaviors without imposing full-game computational costs. Two-Bridge is released as a lightweight, Gym-compatible wrapper on top of PySC2, with maps, wrappers, and reference scripts fully open-sourced to encourage broad adoption as a standard benchmark.