Researchers have introduced CodeCoR, a multi-agent framework for code generation that employs four large language model (LLM) agents to generate prompts, code, test cases, and repair advice. By pruning low-quality outputs at each stage, CodeCoR aims to reduce error propagation in sequential code generation workflows. Experiments on HumanEval and MBPP datasets show that CodeCoR achieves an average Pass@1 score of 77.13%, outperforming existing baselines.
Why it matters: This work demonstrates a significant advance in code generation reliability by integrating self-reflection and output pruning in a multi-agent LLM framework, addressing a key limitation of sequential LLM-based approaches.
AutoSpec is a framework that automatically generates and evaluates neural network specifications for learning-augmented systems, particularly in safety-critical domains. It uses a tree-based algorithm to partition the input space and a statistical certification framework to provide accuracy guarantees for each specification. Experiments across four applications show that AutoSpec improves F1 score by up to 53% over human-defined specifications and 73% over the strongest baseline.
Why it matters: This work automates the specification process in neural network verification, addressing a key bottleneck and making formal safety guarantees more practical for learning-augmented systems.
A large-scale empirical study examines how different vLLM inference engine configurations—specifically attention kernel type, prefix caching, and chunked prefill—affect energy consumption, latency, and accuracy across multiple open-weight LLMs and tasks. The findings show that attention kernel type and prefix caching have significant impacts on energy and performance, while chunked prefill has limited effect under default settings. The optimal configuration is highly dependent on the specific model and workload, with no single configuration performing best in all scenarios.
Why it matters: This research provides practical guidance for optimizing LLM inference deployments, demonstrating that while configuration tuning can yield local improvements, the choice of model has a greater overall impact on trade-offs between energy, performance, and accuracy.
A new empirical study analyzes 255 bug reports from Codex, Gemini-CLI, LangChain, and CrewAI, introducing a new class of bugs termed agent-reactive (AR) bugs that arise at the interface between LLM model outputs and harness code. These AR bugs often manifest as silent errors without clear test oracles, and their reproduction is complicated by the stochastic nature of LLM responses. The study also finds a mismatch between user-proposed fixes, which often suggest harness-side guardrails, and developer responses.
Why it matters: This research exposes a critical challenge in LLM agent reliability, emphasizing the need for new testing and debugging tools to address bugs that emerge from the interaction between model and harness.
BayesContact is a Simulation-Based Inference framework that combines depth and force/torque observations to estimate object pose during peg-in-hole insertion tasks. The method maintains a particle belief over pose, using a renderer and physics simulator to score hypotheses against real observations. In both simulated and real-robot experiments, BayesContact improves pose observability and increases insertion success by 30% compared to vision-only inference.
Why it matters: This work offers a practical approach to fusing vision and tactile sensing for precise pose estimation in contact-rich robotic manipulation, potentially reducing the need for retraining across different environments and geometries.
Researchers have introduced Handroid, a 27-degree-of-freedom (DoF) reconfigurable robot capable of switching between a dexterous hand and a desktop-scale humanoid form (0.33 m tall, 2.05 kg). The platform supports teleoperation, dexterous grasping, in-hand manipulation, humanoid locomotion, and long-horizon tasks that involve changing its embodiment. Handroid is validated through real-world manipulation, reinforcement-learning-based locomotion, and tasks requiring reconfiguration and coordinated actions.
Why it matters: Handroid enables unified research on dexterous manipulation and humanoid mobility within a single, compact, and reconfigurable platform, advancing the field of morphology-reconfigurable robotics and cross-embodiment learning.
Researchers introduce Orbis 2, a hierarchical world model for autonomous driving that separates future prediction into two levels: a high-level predictor for coarse scene structure over long time horizons and a low-level generator for detailed predictions. The model is trained in two stages, first with diffusion forcing pretraining to enhance internal representations, followed by teacher forcing fine-tuning for stable rollouts. Orbis 2 achieves state-of-the-art results on standard driving world model benchmarks, including long-horizon generation fidelity and steering responsiveness.
Why it matters: This work demonstrates a significant advance in autonomous driving world models by combining long-horizon spatial reasoning with high perceptual fidelity, leading to improved performance and internal representations.
A new preprint investigates token-level reduction techniques to lower computational costs in large language model (LLM)-based code summarization. The study finds that optimal token reduction strategies are highly language-dependent: AST-based reduction improves Java summarization performance by 37% but degrades Python by up to 49%, while CrystalBLEU-guided pruning achieves robust cross-language token reduction of 60-72%. Function signature-based reduction is optimal for Python, achieving 83% token reduction while maintaining summary quality.
Why it matters: The work demonstrates that language-aware token curation is crucial for efficient code summarization, challenging assumptions about the transferability of token reduction strategies across programming languages.
NexForge is a framework that automatically synthesizes diverse, executable agent tasks and expert trajectories from high-level capability requirements, removing the need for domain-specific infrastructure. The system generated thousands of terminal and office tasks, significantly improving the performance of Qwen3.5-35B-A3B on benchmarks such as Terminal-Bench 2.0 and GDPval. Scaling up to 43,200 tasks enabled the training of Nex-N2 models, which achieve state-of-the-art open-source results and surpass several proprietary systems.
Why it matters: NexForge demonstrates a scalable approach to expanding LLM agent capabilities by automating task synthesis from requirements, reducing manual engineering and enabling broader domain coverage.
A new preprint introduces DPNeXt, a lightweight multi-scale feature fusion decoder designed to replace the standard Dense Prediction Transformer (DPT) for multi-task dense prediction in robotics. DPNeXt employs dual depthwise separable inverted bottlenecks and a Multi-Task Boundary Guidance strategy to improve efficiency and mitigate negative transfer between tasks. The method reduces trainable parameters by 78.6% compared to DPT and achieves state-of-the-art results on Cityscapes and NYUv2 benchmarks, with faster inference on resource-constrained hardware.
Why it matters: This work advances efficient multi-task perception for robotics, enabling real-time scene understanding on devices with limited computational resources.
SLAC is a reinforcement learning method that uses a low-fidelity simulator to pretrain a task-agnostic latent action space through unsupervised skill discovery, enabling efficient and safe real-world learning for high-degree-of-freedom robots. The approach achieves state-of-the-art performance on bimanual mobile manipulation tasks, learning contact-rich, whole-body behaviors in under an hour of real-world interaction without demonstrations or hand-crafted priors.
Why it matters: SLAC advances the feasibility of real-world reinforcement learning for complex robots by combining simulation-based pretraining with safe, sample-efficient real-world learning.
Researchers introduce a method that focuses dense data collection on the critical, contact-rich segments of manipulation tasks, while using traditional planning for simpler free-space motions. By combining automated data collection with offline deep reinforcement learning, their approach achieves a 96% average success rate across four real-world tasks using only 2–2.5 hours of autonomous data, significantly outperforming the strongest baseline at 55%. The method also maintains high performance in out-of-distribution scenarios where end-to-end approaches typically struggle.
Why it matters: This work shows that targeted data collection can greatly improve the efficiency and generalization of robotic manipulation, reducing reliance on large datasets and teleoperation.
A new prompting technique called MoT (Modularization of Thought) has been proposed to improve code generation by large language models. MoT decomposes complex programming problems into smaller, independent reasoning steps using a Multi-Level Reasoning Graph, enabling a more structured and interpretable approach. Experiments with GPT-4o-mini and DeepSeek-R1 on eight benchmarks show that MoT outperforms existing methods like Chain-of-Thought, with Pass@1 scores between 58.1% and 95.1%.
Why it matters: MoT could enhance the ability of language models to solve complex programming tasks by providing a more modular and interpretable reasoning process.
TARS is an LLM-powered agent integrated into Visual Studio Code that adapts code explanations to a developer's expertise, role, and preferences using a Theory of Mind paradigm. In a controlled experiment with 18 participants, TARS users completed code comprehension tasks 26% faster and reported lower cognitive load compared to a baseline.
Why it matters: This work provides evidence that personalized, in-IDE AI assistance can improve developer productivity and reduce cognitive burden.
Researchers have introduced CLI-Tool-Bench, a new benchmark designed to evaluate large language models' (LLMs) ability to autonomously generate complete command-line interface (CLI) tools from scratch. The benchmark uses black-box differential testing on 94 real-world tasks, requiring models to plan and implement entire repositories without scaffolding. Top-performing LLMs achieved only a 43.8% success rate, underscoring the difficulty of 0-to-1 software generation.
Why it matters: This work highlights a significant gap in current LLM capabilities for fully autonomous software creation, which is crucial for the development of advanced coding agents.
This paper presents an embodied active learning method that adapts object detectors to unknown environments while operating under constraints on both robot navigation time and annotation budget. The approach leverages spatial inconsistency to select informative trajectories and images, targeting failure cases to improve model performance. Experiments in both simulated (AI2-THOR) and real-world (Boston Dynamics Spot robot with YOLOv5) settings demonstrate that the method achieves higher detection accuracy than baseline approaches under the same resource constraints.
Why it matters: The work offers a practical solution for efficiently adapting robotic vision systems to new environments with limited resources, potentially improving real-world deployment of autonomous robots.
Researchers have introduced IMBench, a new benchmark designed to evaluate intuitive manipulation in robots by integrating perception, physical reasoning, action generation, and iterative execution. IMBench features 35 tasks and 14,000 filtered trajectories, requiring models to infer physical structure and generate feasible action sequences under explicit constraints. Initial experiments show that current vision-language and vision-language-action models struggle to combine reasoning with executable plans and to generalize across diverse manipulation scenarios.
Why it matters: IMBench highlights a significant gap in current AI and robotics systems' ability to integrate reasoning with action execution, which is crucial for developing more adaptive and capable robotic manipulation.
Researchers introduce PACE, a framework that enables humanoid robots to dynamically generate and adapt psychologically grounded personas through conversational interaction with users. The system employs an interactive elicitation pipeline to create structured identities, which are then expressed via multimodal behaviors on the Ameca robot. Empirical evaluation demonstrates that dynamically generated personas improve user trust, perceived anthropomorphism, and interaction quality compared to static, generic personas.
Why it matters: This work represents a significant advance in human-robot interaction by enabling robots to personalize their identities in real time, which can enhance trust and engagement in practical applications.
MemoGuard is a lightweight adaptive runtime system that validates episodic memories against safety contracts before reuse in communication-limited robot navigation. In simulation, MemoGuard reduced battery safety violations by 76.6% compared to similarity-only memory reuse and decreased fallback calls by 21.4% relative to always invoking local reasoning. The system was evaluated on an NVIDIA Jetson AGX Xavier and is open-sourced.
Why it matters: MemoGuard advances the safety and efficiency of autonomous robots operating in mission-critical, communication-limited environments by reducing unsafe memory reuse and computational overhead.
SkillNav is a new behavioral skill framework for vision-language model (VLM)-based navigation that leverages curiosity value maps as writable memory, enabling composable skills without retraining. The method introduces a tiered skill intervention system and achieves state-of-the-art SPL on MP3D (25.5), HM3D v0.1 (39.3), and HM3D v0.2 (43.2), with improvements of up to 6.0 SPL over previous methods. SkillNav also attains the highest reported Success Rate on HM3D v0.1 (69.7) and v0.2 (75.9).
Why it matters: SkillNav enables continual, training-free improvement of navigation agents by allowing new skills to be added without retraining, advancing the flexibility and practicality of embodied AI systems.