Counterfactual reasoning requires models to reason beyond the observed world and explain how altered conditions propagate through downstream consequences. Existing benchmarks largely target bounded settings with fixed variables or single gold outcomes, overlooking open-domain scenarios requiring causal-process evaluation. To this end, we present WhatIfBench, a diagnostic benchmark for open-domain, open-form, long-horizon counterfactual causal reasoning, containing 220 what-if questions across STEM, HSS, and Hybrid scenarios. To evaluate free-form responses, we further propose PRISM, which first converts each natural-language explanation into a Response-Derived Semantic Causal Graph of events, states, and mechanisms. On top of this graph, PRISM then jointly applies a Process Metric assessing graph-level causal validity and a Rubric Metric assessing answer-level explanatory adequacy. Evaluating six frontier LLMs with this framework, we find that WhatIfBench remains far from saturated: even the strongest model reaches only a 64.62% final score. Further analysis reveals persistent causal gaps, premise drift, and topology fragmentation, suggesting that fluent counterfactual narratives often mask fragile causal processes. The benchmark, code, and evaluation scripts are available at https://github.com/zju-gt/WhatIfBench.

Multimodal large language models show significant potential in mathematical reasoning, yet they remain constrained by a scarcity of high-quality training data for visually grounded problems. Existing approaches commonly suffer from narrow coverage, image-text inconsistency, and heavy imitation of seed problems. To address these limitations, this paper proposes AgSynth, a multi-agent collaborative reverse synthesis framework. It first constructs solution logic programmatically and renders the corresponding images, then generates problem statements backward from the visual facts, thereby embedding answer correctness and image-text consistency directly into the generation pipeline. Meanwhile, AgSynth distills the core competencies tested in seed problems and performs divergent reconstruction, enabling synthesized problems to be substantially novel in both structure and visual presentation. Based on this framework, we construct the AgSynth-10K dataset, comprising 10,477 multimodal mathematical reasoning problems that have undergone two-tier quality control, spanning four major domains: geometry, algebra, combinatorics, and number theory. Experiments demonstrate that reinforcement learning training on AgSynth-10K significantly improves model performance on reasoning-intensive benchmarks (WeMath +7.42, LogicVista+9.15), validating the effectiveness of this traceable synthesis paradigm. The framework is publicly available at https://github.com/Yaxuan7690/AgSynth.

Large language models (LLMs) show growing promise for scientific discovery, yet existing benchmarks rely on static, single-turn QA formats that cannot measure the iterative reasoning and strategic planning required in real research. We introduce MolQuest, an agent-based benchmark that reformulates molecular structure elucidation as a multi-turn interactive task grounded in authentic experimental data. Within a simulated laboratory, the model must actively plan experiments, acquire heterogeneous spectral evidence (NMR, MS, IR) on demand, and iteratively refine structural hypotheses while minimizing unnecessary tool calls—mirroring the abductive reasoning workflow of a practicing chemist. We evaluate 16 frontier LLM variants under two settings, yielding 32 model–setting configurations, and find that the best Agent-mode result reaches only 51.51% accuracy, while most remain substantially lower. Furthermore, the dynamic paradigm amplifies performance for some models yet degrades others, revealing strategic planning deficiencies invisible to static evaluations. MolQuest provides a reproducible, extensible framework that exposes the critical gap between pattern recognition and genuine scientific reasoning in current LLMs.

Real-world clinical competence demands more than answering medical knowledge questions—it requires proactively gathering incomplete patient information, selecting appropriate examinations, and formulating safe, cost-conscious treatment plans through multi-turn interaction. Existing medical LLM benchmarks rely predominantly on static QA formats, failing to assess interactive information acquisition, tool use, or cost–quality tradeoffs. We propose MedCPEnv, a reproducible, cost-aware simulated hospital environment for evaluating and training medical LLM agents. Built from MIMIC-IV-NOTE discharge notes, MedCPEnv comprises 48,803 curated cases, a 1,000-case benchmark, and a 371,550-entry medical QA knowledge base, exposing four deterministic action types (ASK, EXAM, KNOWLEDGE, FINAL) to eliminate high-variance LLM role-playing artifacts. Evaluating eight frontier models, GPT-5.4 achieves the highest composite score (3.69/5) and diagnostic accuracy (60.3%), yet treatment accuracy remains only 19.4%, identifying treatment plan generation as the primary bottleneck. Substantial inter-model variation in the cost–quality tradeoff further demonstrates that clinical correctness alone is insufficient to characterize agent behavior. GRPO-based Agentic RL on Qwen3-4B shows that medical quality rewards consistently improve tool-use strategies, raising the composite score from 2.97 to 3.14 and diagnostic accuracy from 26.0% to 31.2%, establishing MedCPEnv as both a rigorous evaluation benchmark and a reproducible training environment for medical agent research.

Open-ended medical LLM evaluation remains weakly grounded in physician-calibrated coverage of clinically relevant response criteria, especially in localized clinical settings. We introduce ClinConsensus, a Chinese medical benchmark of 2,500 expert-curated cases spanning 36 specialties, 12 task themes, multiple difficulty levels, and lay-facing versus professional-facing settings. Each case is paired with 30 case-specific binary rubric criteria. To evaluate whether responses satisfy enough physician-authored criteria, we propose Clinician-Anchored Coverage Score (CACS), a physician-calibrated threshold metric instantiated at k = 10, and develop a dual-judge framework combining a GPT-5.1 grader with a physician-supervised Qwen3-8B judge. Evaluating 11 frontier LLMs, we find a persistent coverage gap: Rubric Accuracy ranges from 39.6% to 52.1%, whereas CACS@10 ranges from 17.8% to 32.9%, leaving a 19.2–21.9 point gap across models. Stratified analyses further reveal substantial variation across reasoning, evidence use, structured extraction, medication instructions, follow-up, and dialogue register. These results suggest that medical LLM evaluation should measure thresholded, rubric-grounded clinical coverage rather than average partial correctness.

As code large language models (LLMs) evolve into tool-interactive agents via the Model Context Protocol (MCP), their generalization is increasingly limited by low-quality synthetic data and the diminishing returns of quantity scaling; moreover, quantity-centric scaling exhibits an early bottleneck that underutilizes trajectory data. We propose TDScaling, a Trajectory Diversity Scaling-based data synthesis framework for code agents that scales performance through diversity rather than raw volume. Moreover, TDScaling is more data-efficient: under a fixed training budget, increasing trajectory diversity yields larger gains than adding more trajectories, improving the performance-cost trade-off for agent training. TDScaling integrates four innovations: (1) a Business Cluster mechanism that captures real-service logical dependencies; (2) a Blueprint-driven multi-agent paradigm that enforces trajectory coherence; (3) an adaptive evolution mechanism that steers synthesis toward long-tail scenarios using Domain Entropy, Reasoning Mode Entropy, and Cumulative Action Complexity to prevent mode collapse; and (4) a sandboxed code tool that mitigates catastrophic forgetting of intrinsic coding capabilities. Experiments on general tool-use benchmarks (BFCL, 𝜏²-Bench) and code agent tasks (RebenchT, CodeCI, BIRD) demonstrate a win-win outcome: TDScaling improves both tool-use generalization and inherent coding proficiency. Crucially, we show that trajectory diversity scaling attains a substantially higher performance ceiling than quantity scaling, establishing a resource-efficient paradigm for training robust code agents under data bottlenecks.

Existing Text-to-Image (T2I) benchmarks remain anchored in foundational criteria such as image quality and text-image alignment, failing to capture the nuanced capabilities that professional creators demand in real-world workflows. Meanwhile, evaluation pipelines that rely on an untrained multimodal model as a judge diverge from human standards. We introduce Qwen-Image-Bench, a creator-centric benchmark that extends evaluation with two application-driven dimensions (Real-world Fidelity and Creative Generation), alongside the conventional pillars of Quality, Aesthetics, and Alignment. The benchmark organizes these into a three-level taxonomy (5 pillars, 23 sub-capabilities, 56 verifiable rubrics) and provides 1000 stratified bilingual prompts. We train a unified Judge Model (Q-Judger) on Qwen3.6-27B, supervised by 80 professional annotators under blind labeling and triple-review protocols, that scores every image across all 56 facets with fine-grained, rubric-grounded diagnostics. Empirically, Qwen-Image-Bench reliably distinguishes leading T2I models, achieving the greatest separation on the application-driven dimensions where existing benchmarks provide limited insight.

Lawyer-client consultation is a critical starting point for legal services. Effective legal assistance hinges on eliciting sufficient and truthful information from clients in order to devise strategies that best protect their interests. This task requires Large Language Models (LLMs) not only to perform robust legal reasoning, but also to strategically elicit material facts through multi-turn interactions and effectively guide clients with diverse personalities. Yet existing legal benchmarks overlook this interactive capability. To fill this gap, we introduce DLawBench, a diagnostic benchmark for real-world legal consultation. Drawing on realistic client behavior, we characterize lawyer-client interactions into four types: Cooperative, Dependent, Withdrawn, and Adversarial. Using dialogues grounded in real cases, DLawBench evaluates whether LLMs can effectively conduct legal consultation under realistic conditions. DLawBench comprises 461 cases from Chinese and U.S. law, 5,532 paired fact entries, 3,411 inquiry rubrics, and 3,348 issue-resolution rubrics, and evaluates 26 representative LLMs. Systematic experiments show substantial headroom: the best-performing model, GPT-5.5, achieves only 0.562 on consultation-grounded legal reasoning. More importantly, DLawBench exposes both sycophancy in legal consultation and a paradox: models perform worse when clients need guidance most.

Reward models play an essential role in aligning visual generative models, yet most existing visual reward models use a single scalar score or rely on fixed criteria that cannot adapt to different instructions. This limits both interpretability and task sensitivity, especially for text-to-image generation and instruction-based image editing, where different inputs require different evaluation dimensions. We propose RubricRM, a pairwise generative reward modeling framework that first produces an input-specific rubric with evaluation dimensions, weights, and scoring criteria, and then applies the rubric to score candidate images. We train dedicated RubricRM models for text-to-image generation and image editing using a two-stage training pipeline: supervised fine-tuning teaches the model the rubric-based scoring paradigm, while GRPO further improves scoring through fine-grained dimension-level rewards. Experiments on multiple generation and editing benchmarks show that RubricRM outperforms existing specialized reward models and remains competitive with strong proprietary MLLM judges despite using smaller backbones. Our models, data, and code are available at https://github.com/zijiankan/RubricRM.

We introduce SocialArena, a benchmark for systematically defining and evaluating the social intelligence of large language models (LLMs). Although LLMs are increasingly deployed in socially interactive applications such as mental health counseling, companionship chatbots, and educational tutoring, existing evaluations remain fragmented—focusing on isolated capabilities or single scenarios—and evaluation reliability is largely unverified. This work makes three main contributions to address these limitations. (1) We first propose a unified three-tier framework that defines social intelligence along three capability dimensions: role-playing ability, mental state reasoning, and behavioral reasoning. To operationalize this definition, we design a role-playing dialogue paradigm with explicit character thinking for evaluation, enabling simultaneous assessment of these capabilities in a manner that mirrors realistic human interactions. (2) To ensure diversity across social scenarios and thereby prevent evaluation conclusions from being confined to specific scenarios, we introduce the U-TRIMF framework, which guides us in synthesizing social contexts with broad coverage. (3) To obtain reliable evaluation, we invest 132.4K to recruit and train human annotators and adopt a human-annotated arena-style protocol, where annotators compare model responses and provide preference labels along with natural-language critiques. Experiments on a wide range of frontier LLMs reveal substantial gaps in their social intelligence capabilities. We hope that SocialArena can provide valuable insights and guidance for future LLM improvements.

As large language models (LLMs) are increasingly applied to legal domain-specific tasks, evaluating their ability to perform legal work in real-world settings has become essential. However, existing legal benchmarks rely on simplified and highly standardized tasks, failing to capture the ambiguity, complexity, and reasoning demands of real legal practice. Moreover, prior evaluations often adopt coarse, single-dimensional metrics and do not explicitly assess fine-grained legal reasoning. To address these limitations, we introduce PLawBench, a Practical Law Benchmark designed to evaluate LLMs in realistic legal practice scenarios. Grounded in real-world legal workflows, PLawBench models the core processes of legal practitioners through three task categories: public legal consultation, practical case analysis, and legal document generation. These tasks assess a model’s ability to identify legal issues and key facts, perform structured legal reasoning, and generate legally coherent documents. PLawBench comprises 850 questions across 13 practical legal scenarios, with each question accompanied by expert-designed evaluation rubrics, resulting in approximately 12,500 rubric items for fine-grained assessment. Using an LLM-based evaluator aligned with human expert judgments, we evaluate 10 state-of-the-art LLMs. Experimental results show that none achieves strong performance on PLawBench, revealing substantial limitations in the fine-grained legal reasoning capabilities of current LLMs and highlighting important directions for future evaluation and development of legal LLMs. Data is available at: https://anonymous.4open.science/r/PLawbench-B524/.

In this paper we describe the components of our statistical machine translation system. This system combines phrase-to-phrase translations extracted from a bilingual corpus using different alignment approaches. Special methods to extract and align named entities are used. We show how a manual lexicon can be incorporated into the statistical system in an optimized way. Experiments on Chinese-to-English and Arabic-to-English translation tasks are presented.