AI research papers.

Long-horizon terminal agents may solve a task in one run yet fail across repeated runs.

On-policy distillation (OPD) can succeed with surprisingly few training questions, motivating a view of questions as carriers of distributional knowledge transfer.

Generative reward models (GenRMs) generate natural language rationales to justify preference judgments, yet a correct verdict does not guarantee a sound justification.

Aiming to modify textual content while preserving the original style, Scene Text Editing (STE) has significant practical value in many applications.

Existing reasoning evaluations typically provide task-relevant information to models as available premises, rarely examining concurrently how models discover evidence autonomously and how information presentation within…

As model parameters scale up, sparsely activated Mixture-of-Experts (MoE) approaches have become essential for model scaling.

Looped language models can run more iterations at test time, but additional iterations need not help them solve longer problems.

Feed-forward 3D Gaussian Splatting (FF3DGS) reconstructs 3D scenes from images in a single forward pass, making it well-suited to applications that must operate within limited resource budgets.

Video generation models are promising foundations for embodied world models, but only if their generated rollouts are physically valid and faithful to task instructions.

Vision-language-action (VLA) robustness is commonly summarized by aggregate task success, which obscures how individual executions change under distribution shifts and where failures emerge in long-horizon tasks.

Enterprise agents answer questions over workplace records that evolve through team collaboration.

Real-world observations are noisy and often contain excessive task-irrelevant information.

Moving from controllable videos to genuinely playable worlds requires persistent world state, executable game rules, and low-latency interaction.

Training coding agents typically updates the policy while keeping the harness fixed.

As a prototypical latent world model, Joint-Embedding Predictive Architectures (JEPAs) provide strong semantic representations.

Geometric Relative Positional Encoding has become the primary mechanism for injecting spatial inductive biases into multi-view transformers.

Spatial reasoning asks a model to work out where objects are in 3D space, how they relate to one another, and how they move.

LLM agents are commonly trained with RLVR, but the sparse nature of typical RLVR rewards makes credit assignment challenging, especially for long-horizon tasks.

As streaming video generation advances, enabling real-time interaction with generated video is becoming increasingly important.

Code-based video generation offers fine-grained control, targeted editing, and component reuse compared with direct video generation.

Multimodal rerankers are usually dedicated models: a vision-language model (VLM) is trained on retrieval data to score the candidates of a query.

Large Language Model inference over long contexts is increasingly bottlenecked by the cost of attending to a growing Key-Value (KV) cache.

Omni LLMs can naturally locate key audio and visual evidence and integrate information across modalities for audiovisual reasoning.

People adapt how they talk to the person in front of them.

Decision models support routing, guardrails, and output verification in LLM systems.

Despite the success of video world models, extending them to multiple agents is difficult because it requires generating several agent-specific video streams that are consistent in time and between agents.

Continuous flow language models offer a path to parallel generation with tools for control and acceleration unavailable to discrete diffusion, yet effective techniques for building them at the billion-parameter scale rem…

Large Language Model (LLM) agents can handle increasingly complex tasks, navigating long trajectories of observations and actions.

In this report, we introduce Sage-A1.5, a 35B Mixture-of-Experts agentic model that builds on A1 to advance long-horizon problem solving through adaptive capability composition.

Agentic robotics systems combine planners with robot harnesses to perform complex manipulation tasks.

Reliable multi-view forecasting requires videos from different cameras to describe the same evolving 3D scene.

Computer-use agents are limited by the interfaces humans give them.

Game engines and coding agents can specify the layout, motion, and camera of a scene, yet turning such coarse structure into realistic video requires a renderer that follows the structure and completes the appearance it…

Video generation models can now depict remarkably realistic motion.

Scaling data, diversity and model capacity has improved embodied learning in physical environments, where diverse experiences remain grounded in shared physical laws.

We present Qwen-Audio-3.1-ASR, a family of audio-language models that builds upon its predecesor Qwen-Audio-3.0-ASR with improved speech recognition and expanded capabilities in speaker-diarized transcription and general…

Autonomous agents can increasingly propose ideas, modify code, and run experiments, creating an opportunity to accelerate scientific discovery using teams of collaborative AI researchers.

Selecting intermediate features from a frozen diffusion model for discriminative tasks is non-trivial because the useful signal is distributed across the denoising trajectory, network depth, and channels.

Diffusion language models form answers through a sequence of revisable continuous states, but multi-agent systems usually discard those states and make agents communicate only after decoding text.

Deep research systems iteratively plan, gather evidence, synthesize, and revise, making end-to-end latency a key bottleneck for interactive use.

Voice style conversion (VSC) aims to transform a source utterance to match the timbre, accent, and emotion of a target voice while preserving linguistic content.

We present HyperDrone, an extremely fast quadrotor simulator focused on high-performance batch rendering that is more than 20x faster than existing state-of-the-art batch quadrotor simulators.

We introduce OvisOCR3, a 0.8B model for a broad range of OCR-related tasks.

On-policy distillation can transfer the reasoning capabilities of a strong teacher to a smaller student.

Learn how OpenAI is improving ChatGPT across low-resource programming languages, with new evaluations for software engineering, SQLite, and COBOL migration.

Diffusion language models can write several tokens in one forward pass, and they decide during generation how many passes to spend.

Scene-level 3D generation has long been dominated by 2D multi-view or video diffusion models.

LLM routing dynamically selects and coordinates heterogeneous language models across diverse tasks.

Streaming video world models generate the future chunk by chunk, conditioned on their own past and a stream of actions.

We study language-conditioned local navigation, where a robot must navigate to a text-specified target visible in its initial observation and stop at the described location.