Trending Papers

SkillOpt introduces a systematic text-space optimizer for agent skills that trains skills as external agent state with stable updates and zero deployment inference overhead, achieving superior performance across multiple benchmarks and execution environments.

A multi-agent framework using large language models for stock trading simulates real-world trading firms, improving performance metrics like cumulative returns and Sharpe ratio.

FastContext separates repository exploration from code solving in LLM agents using specialized exploration models that reduce token consumption and improve resolution rates.

Kronos, a specialized pre-training framework for financial K-line data, outperforms existing models in forecasting and synthetic data generation through a unique tokenizer and autoregressive pre-training on a large dataset.

A vision-language model operates continuously in real-time, making autonomous decisions about when to respond or delegate, enabling interactive systems that perceive and act upon environmental changes without user prompting.

A large language model adapted for time-series forecasting achieves near-optimal zero-shot performance on diverse datasets across different time scales and granularities.

PaddleOCR-VL-1.6 enhances document parsing performance through targeted data optimization and progressive post-training techniques, achieving state-of-the-art results on OmniDocBench v1.6.

OpenDevin is a platform for developing AI agents that interact with the world by writing code, using command lines, and browsing the web, with support for multiple agents and evaluation benchmarks.

DreamX-World 1.0 is a interactive text/image-to-video model that generates long-horizon content with camera control and scene persistence using specialized encoding, training techniques, and optimization methods.

MinerU2.5, a 1.2B-parameter document parsing vision-language model, achieves state-of-the-art recognition accuracy with computational efficiency through a coarse-to-fine parsing strategy.

PagedAttention algorithm and vLLM system enhance the throughput of large language models by efficiently managing memory and reducing waste in the key-value cache.

LMCACHE enables efficient KV cache management for large language models by storing caches outside GPU memory, supporting cache reuse across queries and inference engines while achieving significant throughput improvements.

Mem0, a memory-centric architecture with graph-based memory, enhances long-term conversational coherence in LLMs by efficiently extracting, consolidating, and retrieving information, outperforming existing memory systems in terms of accuracy and computational efficiency.

VibeThinker-3B demonstrates that compact models can achieve state-of-the-art performance on verifiable reasoning tasks through specialized training techniques, challenging conventional scaling assumptions.

FaraGen creates synthetic datasets for computer use agents, enabling the training of efficient and high-performing models like Fara-7B on diverse web tasks, outperforming larger models on benchmarks.

We study allowing large language models (LLMs) to process arbitrarily long prompts through the lens of inference-time scaling. We propose Recursive Language Models (RLMs), a general inference strategy that treats long prompts as part of an external environment and allows the LLM to programmatically examine, decompose, and recursively call itself over snippets of the prompt. We find that RLMs successfully handle inputs up to two orders of magnitude beyond model context windows and, even for shorter prompts, dramatically outperform the quality of base LLMs and common long-context scaffolds across four diverse long-context tasks, while having comparable (or cheaper) cost per query.

A self-improving AI framework simultaneously updates both model weights and task-specific agent architecture through a language-model feedback agent across legal classification, GPU optimization, and biological data denoising tasks.

SCAIL-2 enables end-to-end character animation by directly transferring motion from driving videos without intermediate representations, using unified task decomposition and synthetic data generation.

Cosmos 3 is an omnimodal world model that processes and generates multiple data types through a unified mixture-of-transformers architecture, achieving state-of-the-art performance in various understanding and generation tasks.

A geometric action model leverages pretrained geometric foundation models to enable language-conditioned manipulation policies with improved accuracy, robustness, and efficiency in 3D physical environments.

InterleaveThinker enables interleaved generation capabilities for image generators through a multi-agent pipeline with planner and critic agents, achieving performance comparable to state-of-the-art models while enhancing reasoning benchmarks.

SmolDocling is a compact vision-language model that performs end-to-end document conversion with robust performance across various document types using 256M parameters and a new markup format.

An AI framework called Arbor enables autonomous scientific research by combining strategic coordination, isolated hypothesis testing, and a persistent knowledge tree to iteratively improve research outcomes across multiple domains.

EverMemOS presents a self-organizing memory system for large language models that processes dialogue streams into structured memory cells and scenes to enhance long-term interaction capabilities.

ARIS is an open-source research harness that uses cross-model adversarial collaboration to ensure reliable long-term research outcomes through coordinated execution, orchestration, and assurance layers.

An omnimodal agent orchestration framework is presented that enables efficient collaboration across multiple modalities through unified task decomposition and specialized sub-agent execution, achieving superior performance on complex multimodal benchmarks.

A novel GPT-based model, OmniFlatten, enables real-time natural full-duplex spoken dialogue through a multi-stage post-training technique that integrates speech and text without altering the original model's architecture.

Robust-U1 enhances multimodal large language models' robustness against visual corruptions through self-recovery capabilities that improve both visual quality and reasoning performance.

A 2B-parameter continuous autoregressive text-to-speech model trained on a multilingual corpus achieves state-of-the-art performance on multiple benchmarks while enabling efficient low-latency speech generation through specialized distillation techniques.

Enhancements to the AgentScope platform improve scalability, efficiency, and ease of use for large-scale multi-agent simulations through distributed mechanisms, flexible environments, and user-friendly tools.

Agents' Last Exam (ALE) is a benchmark for evaluating AI agents on long-term, economically valuable real-world tasks across 13 industry clusters with 1K+ tasks, revealing significant gaps between benchmark performance and practical deployment.

AgentScope enhances agentic applications by providing flexible tool-based interactions, unified interfaces, and advanced infrastructure based on the ReAct paradigm, supporting efficient and safe development and deployment.

MRAgent combines associative memory graphs with active reconstruction to enable dynamic memory access during reasoning, improving long-horizon memory reasoning while reducing computational costs.

VibeVoice synthesizes long-form multi-speaker speech using next-token diffusion and a highly efficient continuous speech tokenizer, achieving superior performance and fidelity.

Zep, a memory layer service, outperforms MemGPT in the DMR benchmark and LongMemEval by excelling in dynamic knowledge integration and temporal reasoning, critical for enterprise use cases.

LightRAG improves Retrieval-Augmented Generation by integrating graph structures for enhanced contextual awareness and efficient information retrieval, achieving better accuracy and response times.

AI-Trader presents the first fully automated live benchmark for evaluating large language models in financial decision-making across multiple markets with autonomous information processing.

Unified vision-language models treat understanding and generation as integrated processes rather than separate tasks, demonstrating strong performance across multiple multimodal capabilities including image synthesis and action reasoning.

DRPO improves LLM reinforcement learning stability by replacing hard masks with smooth regularization that provides continuous gradient corrections beyond trust-region boundaries.

YOLO26 addresses real-time vision challenges through a unified model family with NMS-free inference, improved training strategies, and multi-task capabilities spanning detection, segmentation, and pose estimation.

Person-grounded AI skills are automatically distilled from heterogeneous traces into inspectable, correctable packages that capture both capabilities and behavioral patterns.

RAG-Anything is a unified framework that enhances multimodal knowledge retrieval by integrating cross-modal relationships and semantic matching, outperforming existing methods on complex benchmarks.

MiniMax Sparse Attention enables efficient processing of ultra-long contexts in large language models through blockwise sparsity and optimized GPU execution, achieving significant speedups while maintaining performance.

DataFlow is an LLM-driven data preparation framework that enhances data quality and reproducibility for various tasks, improving LLM performance with automatically generated pipelines.

RF-DETR, a light-weight detection transformer, uses weight-sharing NAS to optimize accuracy and latency for real-time detection across diverse datasets.

A Hierarchical Recurrent Model architecture with specialized training on instruction-response pairs achieves competitive language modeling performance with significantly reduced computational requirements compared to traditional Transformer-based approaches.

A 20B search agent trained with reinforcement learning within a stateful search framework demonstrates superior retrieval performance across multiple domains by separating semantic decision-making from environmental bookkeeping.

GLM-5 advances foundation models with DSA for cost reduction, asynchronous reinforcement learning for improved alignment, and enhanced coding capabilities for real-world software engineering.

SpatialClaw is a training-free framework that uses code as an action interface to enable flexible, stateful spatial reasoning in vision-language models, achieving superior performance across diverse 3D/4D spatial reasoning tasks.

Memanto presents a universal memory layer for agentic AI that eliminates computational overhead of hybrid semantic graph architectures through a typed semantic memory schema and information-theoretic search engine.