LLM-powered agents struggle with personalization when users issue raw, underspecified queries. In this setting, agents must infer latent intent, extract preferences from noisy interaction histories, and select among competing alternatives. Existing benchmarks rarely test this capability, as they often rely on user-refined queries or simplified histories. We introduce personalized product search (PPS), a testbed for agentic personalization under raw queries and diverse histories. We construct **A**gent **Pe**rsonalized **B**enchmark (**APeB**) from action logs, pairing underspecified intents with rich histories and user-viewed candidate items. Evaluating state-of-the-art LLMs with multi-step agent workflows, we find that models handle explicit queries well but struggle with early-stage queries requiring intent and preference discovery. Rubric analysis attributes this gap mainly to ineffective history use. A simple history-aware query-refinement pipeline, VQRA, yields consistent gains, highlighting the need for dedicated history-utilization modules in personalized agents.

Think-with-images, where multimodal LLMs (MLLMs) actively manipulate visual inputs as part of their reasoning, has become a focal research direction for agentic multimodal systems, with a surge of prompting, supervised, and reinforcement learning (RL) methods teaching models to invoke vision tools such as cropping, zooming, and editing. Yet most established multimodal benchmarks use images whose critical content is already easy to perceive in a single pass, and many have saturated on frontier models. We introduce VisualToolBench (VTB), a benchmark deliberately curated so that critical visual content is dense, occluded, dispersed, rotated, or low-contrast, making active image manipulation a prerequisite for success. VTB contains 1,204 expert-authored, open-ended tasks (603 single-turn, 601 multi-turn) across five domains. Each task is paired with rubrics tagged by five competency categories (Visual Understanding, Truthfulness, Instruction Following, Reasoning, and Writing Style) that enable per-axis disentanglement. Across 26 evaluated MLLMs, even the strongest model reaches only 29% accuracy, with Visual Understanding the consistent bottleneck and visual misinterpretation accounting for 32% to 39% of failures. Top models, moreover, converge to comparable accuracy through markedly different tool-use strategies, which we analyze in detail.

Verification is critical for improving agents: it provides the reward signal for Reinforcement Learning and enables inference-time gains through Test-Time Scaling (TTS). Despite its importance, verification in software engineering (SWE) agent settings often relies on code execution, which can be difficult to scale due to environment setup overhead. Scalable alternatives such as patch classifiers and heuristic methods exist, but they are less grounded in codebase context and harder to interpret. To this end, we explore Agentic Rubrics: an expert agent interacts with the repository to create a context-grounded rubric checklist, and candidate patches are then scored against it without requiring test execution. On SWE-Bench Verified under parallel TTS evaluation, Agentic Rubrics achieve a score of 54.2% on Qwen3-Coder-30B-A3B and 40.6% on Qwen3-32B, with at least a +3.5 percentage-point gain over the strongest baseline in our comparison set. We further analyze rubric behavior, showing that rubric scores are consistent with ground-truth tests while also flagging issues that tests do not capture. Our ablations show that agentic context gathering is essential for producing codebase-specific, unambiguous criteria. Together, these results suggest that Agentic Rubrics provide an efficient, scalable, and granular verification signal for SWE agents.

Frontier model progress is often measured using academic benchmarks that provide a limited view of performance on open-ended, economically consequential tasks in high-stakes professional domains where practical returns matter most. We introduce Professional Reasoning Bench (PRBench), a realistic, open-ended, and difficult benchmark of real-world problems in Finance and Law. We open-source its 1,100 expert-authored tasks and 19,356 expert-curated criteria, making it the largest public, rubric-based benchmark for both legal and finance domains. We recruit 182 qualified professionals, holding JDs, CFAs, or 6+ years of experience, who contributed questions inspired by their actual workflows. This process yields significant diversity, with tasks spanning 114 countries and 47 US jurisdictions. Our expert-curated rubrics are validated through a rigorous quality pipeline, including independent expert validation. Subsequent evaluation of 20 leading models reveals substantial room for improvement, with top scores of only 0.39 (Finance) and 0.37 (Legal) on our Hard subsets. We further catalog associated economic impacts of the prompts and analyze performance using human-annotated rubric categories. Common failure modes include inaccurate judgments, a lack of process transparency and incomplete reasoning, highlighting critical gaps in their reliability for professional adoption.

End-to-end (E2E) spoken dialogue systems are replacing cascaded pipelines for voice-based human-AI interaction. Existing benchmarks primarily evaluate these systems on synthetic speech and single-turn tasks, leaving multi-turn conversational ability underexplored. We introduce Audio MultiChallenge an open-source benchmark to evaluate these systems under natural multi-turn interaction patterns. Building on the text-based MultiChallenge framework, which evaluates Inference Memory, Instruction Retention, and Self Coherence, we introduce a new axis Voice Editing that tests robustness to mid-utterance speech repairs and backtracking. We augment each axis to the audio modality, such as introducing Audio-Cue challenges for Inference Memory that require recalling ambient sounds and paralinguistic signals beyond semantic content. We curate 452 conversations from 47 speakers with 1,712 instance-specific rubrics through a hybrid pipeline that exposes model failures at scale while preserving natural disfluencies found in unscripted human speech. Our evaluation reveals that even frontier models struggle on our benchmark, with our highest-performing model achieving a 54.65% pass rate. Error analysis shows that models are not sufficiently robust to human speech when tracking instructions, edits, and audio cues, highlighting the need for improved audio-native multi-turn interaction capabilities.