EvoStreaming: Your Offline Video Model Is a Natively Streaming Assistant

Streaming video understanding demands more than watching longer videos: assistants must decide when to speak in real time, balancing responsiveness against verbosity. Yet most video-language models (VideoLLMs) are trained for offline inference, and existing streaming benchmarks externalize this timing decision to the evaluator. We address this gap with RealStreamEval, a frame-level multi-turn evaluation protocol that exposes models to sequential observations and penalizes unnecessary responses. Under this protocol, we observed that strong offline VideoLLMs retain useful visual understanding but lack an interaction policy for deciding when to respond. Motivated by this observation, we propose EvoStreaming, a self-evolved streaming adaptation framework in which the base model itself acts as data generator, relevance annotator, and roll-out policy to synthesize streaming trajectories without external supervision. With only $1{,}000$ self-generated samples ($139\times$ less than the leading streaming instruction-tuning approach) and no architectural changes, EvoStreaming consistently improves the overall RealStreamEval score by up to $10.8$ points across five open VideoLLM backbones (Qwen2/2.5/3-VL, InternVL-3.5, MiniCPM-V4.5) while largely preserving offline video performance. These results suggest that data-efficient interaction tuning is a practical path for adapting existing VideoLLMs to streaming assistants.

HeadRouter: Dynamic Head-Weight Routing for Task-Adaptive Audio Token Pruning in Large Audio Language Models

Recent large audio language models (LALMs) demonstrate remarkable capabilities in processing extended multi-modal sequences, yet incur high inference costs. Token compression is an effective method that directly reduces redundant tokens in the sequence. Existing compression methods usually assume that all attention heads in LALMs contribute equally to various audio tasks and calculate token importance by averaging scores across all heads. However, our analysis demonstrates that attention heads exhibit distinct behaviors across diverse audio domains. We further reveal that only a sparse subset of attention heads actively responds to audio, with completely different performance when handling semantic and acoustic tasks. In light of this observation, we propose HeadRouter, a head-importance-aware token pruning method that perceives the varying importance of attention heads in different audio tasks to maximize the retention of crucial tokens. HeadRouter is training-free and can be applied to various LALMs. Extensive experiments on the AudioMarathon and MMAU-Pro benchmarks demonstrate that HeadRouter achieves state-of-the-art compression performance, exceeding the baseline model even when retaining 70% of the audio tokens and achieving 101.8% and 103.0% of the vanilla average on Qwen2.5-Omni-3B and Qwen2.5-Omni-7B, respectively.

STDec: Spatio-Temporal Stability Guided Decoding for dLLMs

Diffusion Large Language Models (dLLMs) have achieved rapid progress, viewed as a promising alternative to the autoregressive paradigm. However, most dLLM decoders still adopt a global confidence threshold, and do not explicitly model local context from neighboring decoded states or temporal consistency of predicted token IDs across steps. To address this issue, we propose a simple spatio-temporal stability guided decoding approach, named STDec. We observe strong spatio-temporal stability in dLLM decoding: newly decoded tokens tend to lie near decoded neighbors, and their predicted IDs often remain consistent across several denoising steps. Inspired by this stability, our STDec includes spatial-aware decoding and temporal-aware decoding. The spatial-aware decoding dynamically generates the token-adaptive threshold by aggregating the decoded states of nearby tokens. The temporal-aware decoding relaxes the decoding thresholds for tokens whose predicted token IDs remain consistent over denoising steps. Our STDec is training-free and remains compatible with cache-based acceleration methods. Across textual reasoning and multimodal understanding benchmarks, STDec substantially improves throughput while maintaining comparable task performance score. Notably, on MBPP with LLaDA, STDec achieves up to 14.17x speedup with a comparable score. Homepage: https://yzchen02.github.io/STDec.

V-CAST: Video Curvature-Aware Spatio-Temporal Pruning for Efficient Video Large Language Models

Video large language models (VideoLLMs) show strong capability in video understanding, yet long-context inference is still dominated by massive redundant visual tokens in the prefill stage. We revisit token compression for VideoLLMs under a tight budget and identify a key bottleneck, namely insufficient spatio-temporal information coverage. Existing methods often introduce discontinuous coverage through coarse per-frame allocation or scene segmentation, and token merging can further misalign spatio-temporal coordinates under MRoPE-style discrete (t,h,w) bindings. To address these issues, we propose V-CAST (Video Curvature-Aware Spatio-Temporal Pruning), a training-free, plug-and-play pruning policy for long-context video inference. V-CAST casts token compression as a trajectory approximation problem and introduces a curvature-guided temporal allocation module that routes per-frame token budgets to semantic turns and event boundaries. It further adopts a dual-anchor spatial selection mechanism that preserves high-entropy visual evidence without attention intervention, while keeping retained tokens at their original coordinates to maintain positional alignment. Extensive experiments across multiple VideoLLMs of different architectures and scales demonstrate that V-CAST achieves 98.6% of the original performance, outperforms the second-best method by +1.1% on average, and reduces peak memory and total latency to 86.7% and 86.4% of vanilla Qwen3-VL-8B-Instruct.

Bridging Visual Representation and Reinforcement Learning from Verifiable Rewards in Large Vision-Language Models

Reinforcement Learning from Verifiable Rewards (RLVR) has substantially enhanced the reasoning capabilities of large language models in abstract reasoning tasks. However, its application to Large Vision-Language Models (LVLMs) remains constrained by a structural representational bottleneck. Existing approaches generally lack explicit modeling and effective utilization of visual information, preventing visual representations from being tightly coupled with the reinforcement learning optimization process and thereby limiting further improvements in multimodal reasoning performance. To address this limitation, we propose KAWHI (Key-Region Aligned Weighted Harmonic Incentive), a plug-and-play reward reweighting mechanism that explicitly incorporates structured visual information into uniform reward policy optimization methods (e.g., GRPO and GSPO). The method adaptively localizes semantically salient regions through hierarchical geometric aggregation, identifies vision-critical attention heads via structured attribution, and performs paragraph-level credit reallocation to align spatial visual evidence with semantically decisive reasoning steps. Extensive empirical evaluations on diverse reasoning benchmarks substantiate KAWHI as a general-purpose enhancement module, consistently improving the performance of various uniform reward optimization methods. Project page: KAWHI (https://kawhiiiileo.github.io/KAWHI_PAGE/)

Flash-Unified: A Training-Free and Task-Aware Acceleration Framework for Native Unified Models

Native unified multimodal models, which integrate both generative and understanding capabilities, face substantial computational overhead that hinders their real-world deployment. Existing acceleration techniques typically employ a static, monolithic strategy, ignoring the fundamental divergence in computational profiles between iterative generation tasks (e.g., image generation) and single-pass understanding tasks (e.g., VQA). In this work, we present the first systematic analysis of unified models, revealing pronounced parameter specialization, where distinct neuron sets are critical for each task. This implies that, at the parameter level, unified models have implicitly internalized separate inference pathways for generation and understanding within a single architecture. Based on these insights, we introduce a training-free and task-aware acceleration framework, FlashU, that tailors optimization to each task's demands. Across both tasks, we introduce Task-Specific Network Pruning and Dynamic Layer Skipping, aiming to eliminate inter-layer and task-specific redundancy. For visual generation, we implement a time-varying control signal for the guidance scale and a temporal approximation for the diffusion head via Diffusion Head Cache. For multimodal understanding, building upon the pruned model, we introduce Dynamic Token Pruning via a V-Norm Proxy to exploit the spatial redundancy of visual inputs. Extensive experiments on Show-o2 demonstrate that FlashU achieves 1.78$\times$ to 2.01$\times$ inference acceleration across both understanding and generation tasks while maintaining SOTA performance, outperforming competing unified models and validating our task-aware acceleration paradigm. Our code is publicly available at https://github.com/Rirayh/FlashU.

OmniSIFT: Modality-Asymmetric Token Compression for Efficient Omni-modal Large Language Models

Omni-modal Large Language Models (Omni-LLMs) have demonstrated strong capabilities in audio-video understanding tasks. However, their reliance on long multimodal token sequences leads to substantial computational overhead. Despite this challenge, token compression methods designed for Omni-LLMs remain limited. To bridge this gap, we propose OmniSIFT (Omni-modal Spatio-temporal Informed Fine-grained Token compression), a modality-asymmetric token compression framework tailored for Omni-LLMs. Specifically, OmniSIFT adopts a two-stage compression strategy: (i) a spatio-temporal video pruning module that removes video redundancy arising from both intra-frame structure and inter-frame overlap, and (ii) a vision-guided audio selection module that filters audio tokens. The entire framework is optimized end-to-end via a differentiable straight-through estimator. Extensive experiments on five representative benchmarks demonstrate the efficacy and robustness of OmniSIFT. Notably, for Qwen2.5-Omni-7B, OmniSIFT introduces only 4.85M parameters while maintaining lower latency than training-free baselines such as OmniZip. With merely 25% of the original token context, OmniSIFT consistently outperforms all compression baselines and even surpasses the performance of the full-token model on several tasks.

Structure-based RNA Design by Step-wise Optimization of Latent Diffusion Model

RNA inverse folding, designing sequences to form specific 3D structures, is critical for therapeutics, gene regulation, and synthetic biology. Current methods, focused on sequence recovery, struggle to address structural objectives like secondary structure consistency (SS), minimum free energy (MFE), and local distance difference test (LDDT), leading to suboptimal structural accuracy. To tackle this, we propose a reinforcement learning (RL) framework integrated with a latent diffusion model (LDM). Drawing inspiration from the success of diffusion models in RNA inverse folding, which adeptly model complex sequence-structure interactions, we develop an LDM incorporating pre-trained RNA-FM embeddings from a large-scale RNA model. These embeddings capture co-evolutionary patterns, markedly improving sequence recovery accuracy. However, existing approaches, including diffusion-based methods, cannot effectively handle non-differentiable structural objectives. By contrast, RL excels in this task by using policy-driven reward optimization to navigate complex, non-gradient-based objectives, offering a significant advantage over traditional methods. In summary, we propose the Step-wise Optimization of Latent Diffusion Model (SOLD), a novel RL framework that optimizes single-step noise without sampling the full diffusion trajectory, achieving efficient refinement of multiple structural objectives. Experimental results demonstrate SOLD surpasses its LDM baseline and state-of-the-art methods across all metrics, establishing a robust framework for RNA inverse folding with profound implications for biotechnological and therapeutic applications.

IPCV: Information-Preserving Compression for MLLM Visual Encoders

Multimodal Large Language Models (MLLMs) deliver strong vision-language performance but at high computational cost, driven by numerous visual tokens processed by the Vision Transformer (ViT) encoder. Existing token pruning strategies are inadequate: LLM-stage token pruning overlooks the ViT's overhead, while conventional ViT token pruning, without language guidance, risks discarding textually critical visual cues and introduces feature distortions amplified by the ViT's bidirectional attention. To meet these challenges, we propose IPCV, a training-free, information-preserving compression framework for MLLM visual encoders. IPCV enables aggressive token pruning inside the ViT via Neighbor-Guided Reconstruction (NGR) that temporarily reconstructs pruned tokens to participate in attention with minimal overhead, then fully restores them before passing to the LLM. Besides, we introduce Attention Stabilization (AS) to further alleviate the negative influence from token pruning by approximating the K/V of pruned tokens. It can be directly applied to previous LLM-side token pruning methods to enhance their performance. Extensive experiments show that IPCV substantially reduces end-to-end computation and outperforms state-of-the-art training-free token compression methods across diverse image and video benchmarks. Our code is available at https://github.com/Perkzi/IPCV.

Mixing Importance with Diversity: Joint Optimization for KV Cache Compression in Large Vision-Language Models

Recent large vision-language models (LVLMs) demonstrate remarkable capabilities in processing extended multi-modal sequences, yet the resulting key-value (KV) cache expansion creates a critical memory bottleneck that fundamentally limits deployment scalability. While existing KV cache compression methods focus on retaining high-importance KV pairs to minimize storage, they often overlook the modality-specific semantic redundancy patterns that emerge distinctively in multi-modal KV caches. In this work, we first analyze how, beyond simple importance, the KV cache in LVLMs exhibits varying levels of redundancy across attention heads. We show that relying solely on importance can only cover a subset of the full KV cache information distribution, leading to potential loss of semantic coverage. To address this, we propose \texttt{MixKV}, a novel method that mixes importance with diversity for optimized KV cache compression in LVLMs. \texttt{MixKV} adapts to head-wise semantic redundancy, selectively balancing diversity and importance when compressing KV pairs. Extensive experiments demonstrate that \texttt{MixKV} consistently enhances existing methods across multiple LVLMs. Under extreme compression (budget=64), \texttt{MixKV} improves baseline methods by an average of \textbf{5.1\%} across five multi-modal understanding benchmarks and achieves remarkable gains of \textbf{8.0\%} and \textbf{9.0\%} for SnapKV and AdaKV on GUI grounding tasks, all while maintaining comparable inference efficiency. Furthermore, \texttt{MixKV} extends seamlessly to LLMs with comparable performance gains. Our code is available at \href{https://github.com/xuyang-liu16/MixKV}{\textcolor{citeblue}{https://github.com/xuyang-liu16/MixKV}}.