ESAM reduces sensitivity to anti-HER2 therapy in HER2-positive breast cancer by activating the mTOR pathway

Cell Death Discovery nature.com science feeds

By: Xiaoting Chen · Xujie Zhou · Yuting Sang · Zehao Wang · Yue Zhou · Zhishuang Gao · Yanhui Wang · Yangsiyuan Zhao · Rui Xu · Bingqiu Xiu · Jingyan Xue · Jiong Wu

16 September 2026 at 08:00

Cell Death Discovery, Published online: 16 September 2026; doi:10.1038/s41420-026-03349-8

ESAM reduces sensitivity to anti-HER2 therapy in HER2-positive breast cancer by activating the mTOR pathway

Multi-omics and spatial transcriptomics reveal that S100A10 drives CD8+ T-cell exhaustion and immune evasion in hepatocellular carcinoma through cPLA2-5-LOX-mediated arachidonic acid metabolism and ferroptosis

Omics in Hepatocellular

By: Jingxu Zhang · Ganggang Wang · Zhijie Zhou · Bingbing Pu · Xin Zhang · Xinyu Ji · Xiaoliang Wang

13 September 2026 at 18:00

Int Immunopharmacol. 2026 Sep 13;189:117355. doi: 10.1016/j.intimp.2026.117355. Online ahead of print.

ABSTRACT

Immune evasion in hepatocellular carcinoma (HCC) represents a major biological barrier limiting the efficacy of immunotherapy, yet its molecular basis remains incompletely understood. Increasing evidence indicates that tumor metabolic reprogramming and ferroptosis-related signaling play critical roles in shaping an immunosuppressive tumor microenvironment (TME); however, the specific regulatory factors involved remain unclear. This study aims to systematically elucidate the functional role of S100 calcium-binding protein A10 (S100A10) in immune evasion in HCC, with a particular focus on the molecular mechanisms by which S100A10 regulates CD8⁺ T-cell exhaustion through arachidonic acid (AA) metabolism and ferroptosis, as well as its potential therapeutic implications. To this end, data from The Cancer Genome Atlas Liver Hepatocellular Carcinoma (TCGA-LIHC) cohort are integrated to analyze the expression patterns of S100A10, its prognostic value, and its association with the immune microenvironment. S100A10 overexpression and knockout models are established in HCCLM3 and MHCC97L cell lines, and S100A10-mediated metabolic pathway reprogramming is characterized using transcriptomic profiling, untargeted metabolomics, and ferroptosis-related functional assays. In parallel, single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics are employed to delineate the cell-type specificity and spatial distribution of S100A10. Furthermore, human CD8⁺ T-cell co-culture systems and orthotopic mouse HCC models are used to evaluate the impact of S100A10 on immune function and responsiveness to anti-programmed cell death protein 1 (anti-PD-1) therapy. The results demonstrate that S100A10 is significantly upregulated in HCC and is closely associated with poor prognosis and an immunosuppressive state. Mechanistically, S100A10 activates cytosolic phospholipase A₂-arachidonate 5-lipoxygenase (cPLA₂-5-LOX)-mediated AA oxidative metabolism, leading to the accumulation of lipid peroxidation products and ferroptosis-associated signals, thereby driving CD8⁺ T-cell exhaustion and promoting immune evasion. Significantly, inhibition of S100A10 reshapes the tumor immune microenvironment (TIME) and enhances the therapeutic efficacy of anti-PD-1 treatment. Collectively, these findings identify S100A10 as a critical regulator of metabolic-immune coupling in HCC and provide a theoretical basis for combinatorial strategies targeting metabolism and immunotherapy.

PMID:42732672 | DOI:10.1016/j.intimp.2026.117355

Multi-omics and spatial transcriptomics reveal that S100A10 drives CD8+ T-cell exhaustion and immune evasion in hepatocellular carcinoma through cPLA2-5-LOX-mediated arachidonic acid metabolism and ferroptosis

(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)

By: Jingxu Zhang · Ganggang Wang · Zhijie Zhou · Bingbing Pu · Xin Zhang · Xinyu Ji · Xiaoliang Wang

13 September 2026 at 18:00

Int Immunopharmacol. 2026 Sep 13;189:117355. doi: 10.1016/j.intimp.2026.117355. Online ahead of print.

ABSTRACT

Immune evasion in hepatocellular carcinoma (HCC) represents a major biological barrier limiting the efficacy of immunotherapy, yet its molecular basis remains incompletely understood. Increasing evidence indicates that tumor metabolic reprogramming and ferroptosis-related signaling play critical roles in shaping an immunosuppressive tumor microenvironment (TME); however, the specific regulatory factors involved remain unclear. This study aims to systematically elucidate the functional role of S100 calcium-binding protein A10 (S100A10) in immune evasion in HCC, with a particular focus on the molecular mechanisms by which S100A10 regulates CD8⁺ T-cell exhaustion through arachidonic acid (AA) metabolism and ferroptosis, as well as its potential therapeutic implications. To this end, data from The Cancer Genome Atlas Liver Hepatocellular Carcinoma (TCGA-LIHC) cohort are integrated to analyze the expression patterns of S100A10, its prognostic value, and its association with the immune microenvironment. S100A10 overexpression and knockout models are established in HCCLM3 and MHCC97L cell lines, and S100A10-mediated metabolic pathway reprogramming is characterized using transcriptomic profiling, untargeted metabolomics, and ferroptosis-related functional assays. In parallel, single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics are employed to delineate the cell-type specificity and spatial distribution of S100A10. Furthermore, human CD8⁺ T-cell co-culture systems and orthotopic mouse HCC models are used to evaluate the impact of S100A10 on immune function and responsiveness to anti-programmed cell death protein 1 (anti-PD-1) therapy. The results demonstrate that S100A10 is significantly upregulated in HCC and is closely associated with poor prognosis and an immunosuppressive state. Mechanistically, S100A10 activates cytosolic phospholipase A₂-arachidonate 5-lipoxygenase (cPLA₂-5-LOX)-mediated AA oxidative metabolism, leading to the accumulation of lipid peroxidation products and ferroptosis-associated signals, thereby driving CD8⁺ T-cell exhaustion and promoting immune evasion. Significantly, inhibition of S100A10 reshapes the tumor immune microenvironment (TIME) and enhances the therapeutic efficacy of anti-PD-1 treatment. Collectively, these findings identify S100A10 as a critical regulator of metabolic-immune coupling in HCC and provide a theoretical basis for combinatorial strategies targeting metabolism and immunotherapy.

PMID:42732672 | DOI:10.1016/j.intimp.2026.117355

An operational perturbation proteomics-based virtual cell model

Nature - Issue - nature.com science feeds

By: Rui Sun · Liujia Qian · Yongge Li · Tong Liu · Honghan Cheng · Xuedong Zhang · Xueya Zhou · Yuecheng Zhan · Guangmei Zhang · Zhengchao Luo · Kunpeng Ma · Chunlong Wu · Dongchen Ji · Zhangzhi Xue · Hongxue Meng · Yuhang Xiang · Dingwei Lei · Qianhe Zhou · Wenbin Hu · Yuhan Deng · Lingling Tan · Qi Xiao · Zhiwei Liu · Lei Zeng · Liqin Qian · Xuan Zheng · Qiong Hu · Nianzi Luo · Weinan E · Peijie Zhou · Han Wen · Yi Zhu · Tiannan Guo

9 September 2026 at 08:00

Nature, Published online: 09 September 2026; doi:10.1038/s41586-026-11001-9

Temporal protein-abundance measurements from systematically perturbed breast cancer cell lines were generated to develop ProteinTalks, a virtual cell model that functions as an operational tool for diverse drug discovery tasks.

DisDop: Distillation with Domain Priors for Open-Vocabulary Aerial Object Detection

cs.AI, q-bio.NC updates on arXiv.org

By: Ruihao Xu · Yong Liu · Yansong Tang · Sule Bai · Xubing Ye · Bingyao Yu · Yutao Guo · Jiwen Lu · Jie Zhou

26 May 2026 at 12:00

arXiv:2605.24639v1 Announce Type: cross Abstract: With the widespread application of drones in recent years, object detection of aerial images has attracted increasing attention, especially open-vocabulary aerial detection which is not restricted to predefined categories. Due to the scarcity of drone's viewpoint images and their significant differences from natural images, it is difficult to achieve satisfying results by directly applying vanilla open-vocabulary detection methods designed for natural scenarios. Some studies propose to transfer knowledge from pre-trained models by using lightweight networks or generating pseudo labels, but they tend to rely on models trained on natural images, neglecting the potential of foundation models specifically tailored for remote sensing and aerial imagery. To address this limitation, we propose DisDop, a unified framework that systematically distills multi-level domain priors from remote sensing foundation models (e.g., RemoteCLIP and DINOv3) into a lightweight detector. Specifically, we first distill visual priors through a teacher fusion strategy that combines RemoteCLIP's cross-modal alignment capability with DINOv3's fine-grained local feature extraction ability, transferring their complementary strengths to the detector's backbone. Second, we distill textual priors embedded in RemoteCLIP's text encoder by explicitly modeling inter-category semantic relationships, while incorporating global contextual priors to enhance local feature representation for small objects. Through this multi-level prior distillation framework, our DisDop achieves new state-of-the-art performance on open-vocabulary aerial detection benchmarks. Extensive ablation analysis also demonstrates the rationality and effectiveness of our proposed modules.

ATIC Promotes LIHC Progression and Serves as an Independent Prognostic Marker: A Pan-cancer Transcriptomic Analysis

Omics in Hepatocellular

By: Hui Li · Wentao Zhang · Xiaojie Zhou · Weiming Wu · Fuwen Yao · Haojie Zhong · Naiyang Zhan · JianWei Chen · Yongqiang Zhan · Xinping Yang

19 May 2026 at 18:00

Curr Mol Med. 2026 May 11. doi: 10.2174/0115665240438824260113042223. Online ahead of print.

ABSTRACT

BACKGROUND: 5-aminoimidazole-4-carboxamide ribonucleotide formyltransferase/ IMP cyclohydrolase(ATIC) is a 64-kDa bifunctional enzyme, 5-aminoimidazole- 4-carboxamide ribonucleotide formyltransferase (AICART) and IMP cyclohydrolase, respectively. catalyzes the last two steps of the purine ab initio biosynthetic pathway. ATIC has been implicated in cancer progression, but its pan-cancer profile and specific prognostic utility in liver hepatocellular carcinoma (LIHC) remain incompletely defined.

METHODS: We analyzed TCGA RNA-seq data across 33 tumor types to assess ATIC expression, diagnostic performance (ROC/AUC), and prognostic associations (OS, DSS, PFI). We correlated ATIC expression with immune infiltration, TMB, MSI, and predicted neoantigen load, and constructed a LIHC-specific prognostic nomogram integrating ATIC and clinicopathologic features. Enrichment analyses (STRING, GO/KEGG, GSEA) and pharmacogenomic correlations (GDSC, CTRP) were performed to explore mechanisms and drug sensitivities.

RESULTS: ATIC was significantly upregulated in 16 tumor types, including LIHC (p<0.001). Pan-cancer ROC analyses showed high diagnostic accuracy in several cancers (examples: CHOL AUC=1.000, LIHC AUC=0.936, LUAD AUC=0.947). High ATIC expression associated with poorer OS in ACC, HNSC, LIHC, and PAAD (eg, LIHC: HR=1.39(1.04-1.85), p=0.028). In LIHC, ATIC correlated with advanced T stage, higher grade, elevated AFP, and shorter OS. Multivariable Cox regression identified ATIC expression and pathological T stage as independent predictors; time-dependent ROC for the LIHC nomogram showed AUCs of 0.711, 0.649, and 0.653 at 1, 3, and 5 years, respectively. GSEA indicated enrichment of PI3K-AKT-mTOR, MYC targets, and cell-cycle pathways in ATIC-high LIHC. High ATIC expression correlated with predicted increased sensitivity to sorafenib, doxorubicin, cisplatin, epothilone, and mitomycin in the TCGA-LIHC cohort.

DISCUSSION: ATIC upregulation across cancers links to tumor progression, immune modulation, and prognosis (LIHC), suggesting oncogenic roles in pan-cancer contexts. TCGA multi-omics show ATIC associates with immune/molecular subtypes, MSI/TMB/neoantigens, and predicts drug sensitivity, indicating diagnostic/prognostic potential.

CONCLUSION: ATIC is broadly upregulated across cancers and functions as an independent prognostic biomarker in LIHC. The ATIC-integrated nomogram shows modest predictive accuracy for LIHC survival. Our results implicate ATIC in oncogenic signaling (PI3K-AKT-mTOR, MYC, and cell-cycle) and suggest ATIC as a candidate biomarker to guide targeted and chemotherapeutic strategies in LIHC. Further in vitro and in vivo validation is warranted.

PMID:42152649 | DOI:10.2174/0115665240438824260113042223

From Concept to Practice: an Automated LLM-aided UVM Machine for RTL Verification

cs.AI, q-bio.NC updates on arXiv.org

By: Junhao Ye · Yuchen Hu · Ke Xu · Dingrong Pan · Qichun Chen · Jie Zhou · Shuai Zhao · Xinwei Fang · Xi Wang · Nan Guan · Zhe Jiang

7 April 2026 at 12:00

arXiv:2504.19959v4 Announce Type: cross Abstract: Verification presents a major bottleneck in Integrated Circuit (IC) development, consuming nearly 70% of the total development effort. While the Universal Verification Methodology (UVM) is widely used in industry to improve verification efficiency through structured and reusable testbenches, constructing these testbenches and generating sufficient stimuli remain challenging. These challenges arise from the considerable manual coding effort required, repetitive manual execution of multiple EDA tools, and the need for in-depth domain expertise to navigate complex designs.Here, we present UVM^2, an automated verification framework that leverages Large Language Models (LLMs) to generate UVM testbenches and iteratively refine them using coverage feedback, significantly reducing manual effort while maintaining rigorous verification standards.To evaluate UVM^2, we introduce a benchmark suite comprising Register Transfer Level (RTL) designs of up to 1.6K lines of code.The results show that UVM^2 reduces testbench setup time by up to UVM^2 compared to experienced engineers, and achieve average code and function coverage of 87.44% and 89.58%, outperforming state-of-the-art solutions by 20.96% and 23.51%, respectively.

HISA: Efficient Hierarchical Indexing for Fine-Grained Sparse Attention

cs.AI, q-bio.NC updates on arXiv.org

By: Yufei Xu · Fanxu Meng · Fan Jiang · Yuxuan Wang · Ruijie Zhou · Zhaohui Wang · Jiexi Wu · Zhixin Pan · Xiaojuan Tang · Wenjie Pei · Tongxuan Liu · Di Yin · Xing Sun · Muhan Zhang

7 April 2026 at 12:00

arXiv:2603.28458v3 Announce Type: replace-cross Abstract: Token-level sparse attention mechanisms, exemplified by DeepSeek Sparse Attention (DSA), achieve fine-grained key selection by scoring every historical key for each query through a lightweight indexer, then computing attention only on the selected subset. While the downstream sparse attention itself scales favorably, the indexer must still scan the entire prefix for every query, introducing an per-layer bottleneck that grows prohibitively with context length. We propose HISA (Hierarchical Indexed Sparse Attention), a plug-and-play replacement for the indexer that rewrites the search path from a flat token scan into a two-stage hierarchical procedure: (1) a block-level coarse filtering stage that scores pooled block representations to discard irrelevant regions, followed by (2) a token-level refinement stage that applies the original indexer exclusively within the retained candidate blocks. HISA preserves the identical token-level top-sparse pattern consumed by the downstream Sparse MLA operator and requires no additional training. On kernel-level benchmarks, HISA achieves up to speedup at 64K context. On Needle-in-a-Haystack and LongBench, we directly replace the indexer in DeepSeek-V3.2 and GLM-5 with our HISA indexer, without any finetuning. HISA closely matches the original DSA in quality, while substantially outperforming block-sparse baselines.

Can Heterogeneous Language Models Be Fused?

cs.AI, q-bio.NC updates on arXiv.org

By: Shilian Chen · Jie Zhou · Qin Chen · Wen Wu · Xin Li · Qi Feng · Liang He

3 April 2026 at 12:00

arXiv:2604.01674v1 Announce Type: new Abstract: Model merging aims to integrate multiple expert models into a single model that inherits their complementary strengths without incurring the inference-time cost of ensembling. Recent progress has shown that merging can be highly effective when all source models are \emph{homogeneous}, i.e., derived from the same pretrained backbone and therefore share aligned parameter coordinates or compatible task vectors. Yet this assumption is increasingly unrealistic in open model ecosystems, where useful experts are often built on different families such as Llama, Qwen, and Mistral. In such \emph{heterogeneous} settings, direct weight-space fusion becomes ill-posed due to architectural mismatch, latent basis misalignment, and amplified cross-source conflict. We address this problem with \texttt{HeteroFusion} for heterogeneous language model fusion, which consists of two key components: topology-based alignment that transfers knowledge across heterogeneous backbones by matching functional module structures instead of raw tensor coordinates, and conflict-aware denoising that suppresses incompatible or noisy transfer signals during fusion. We further provide analytical justification showing that preserving the target adapter basis while predicting structured updates leads to a stable and well-conditioned transfer process. Across heterogeneous transfer, multi-source fusion, noisy-source robustness, and cross-family generalization settings, \texttt{HeteroFusion} consistently outperforms strong merging, fusion, and ensemble baselines.

Attention at Rest Stays at Rest: Breaking Visual Inertia for Cognitive Hallucination Mitigation

cs.AI, q-bio.NC updates on arXiv.org

By: Boyang Gong · Yu Zheng · Fanye Kong · Jie Zhou · Jiwen Lu

3 April 2026 at 12:00

arXiv:2604.01989v1 Announce Type: cross Abstract: Like a body at rest that stays at rest, we find that visual attention in multimodal large language models (MLLMs) exhibits pronounced inertia, remaining largely static once settled during early decoding steps and failing to support the compositional understanding required for cognitive inference. While existing hallucination mitigation methods mainly target perceptual hallucinations concerning object existence or attributes, they remain inadequate for such cognitive hallucinations that require inter-object relational deduction. Through token-wise attention analysis, we identify this visual inertia as a key factor: attention to semantically critical regions remains persistently focused and fails to dynamically support relational inference. We thereby propose a training-free Inertia-aware Visual Excitation (IVE) method that breaks this inertial pattern by modeling cognitive inference as the dynamic responsiveness of visual attention. Specifically, IVE selects visual tokens that are dynamically emerging relative to historical attention trends while distinguishing tokens exhibiting inertial behavior. To further facilitate compositional inference, IVE introduces an inertia-aware penalty that discourages over-concentration and limits the persistence of attention within localized regions. Extensive experiments show that IVE is effective across various base MLLMs and multiple hallucination benchmarks, particularly for cognitive hallucinations.

PsychAgent: An Experience-Driven Lifelong Learning Agent for Self-Evolving Psychological Counselor

cs.AI, q-bio.NC updates on arXiv.org

By: Yutao Yang · Junsong Li · Qianjun Pan · Jie Zhou · Kai Chen · Qin Chen · Jingyuan Zhao · Ningning Zhou · Xin Li · Liang He

3 April 2026 at 12:00

arXiv:2604.00931v2 Announce Type: replace Abstract: Existing methods for AI psychological counselors predominantly rely on supervised fine-tuning using static dialogue datasets. However, this contrasts with human experts, who continuously refine their proficiency through clinical practice and accumulated experience. To bridge this gap, we propose an Experience-Driven Lifelong Learning Agent (\texttt{PsychAgent}) for psychological counseling. First, we establish a Memory-Augmented Planning Engine tailored for longitudinal multi-session interactions, which ensures therapeutic continuity through persistent memory and strategic planning. Second, to support self-evolution, we design a Skill Evolution Engine that extracts new practice-grounded skills from historical counseling trajectories. Finally, we introduce a Reinforced Internalization Engine that integrates the evolved skills into the model via rejection fine-tuning, aiming to improve performance across diverse scenarios. Comparative analysis shows that our approach achieves higher scores than strong general LLMs (e.g., GPT-5.4, Gemini-3) and domain-specific baselines across all reported evaluation dimensions. These results suggest that lifelong learning can improve the consistency and overall quality of multi-session counseling responses.

WFR-FM: Simulation-Free Dynamic Unbalanced Optimal Transport

cs.AI, q-bio.NC updates on arXiv.org

By: Qiangwei Peng · Zihan Wang · Junda Ying · Yuhao Sun · Qing Nie · Lei Zhang · Tiejun Li · Peijie Zhou

3 April 2026 at 12:00

arXiv:2601.06810v2 Announce Type: replace-cross Abstract: The Wasserstein-Fisher-Rao (WFR) metric extends dynamic optimal transport (OT) by coupling displacement with change of mass, providing a principled geometry for modeling unbalanced snapshot dynamics. Existing WFR solvers, however, are often unstable, computationally expensive, and difficult to scale. Here we introduce WFR Flow Matching (WFR-FM), a simulation-free training algorithm that unifies flow matching with dynamic unbalanced OT. Unlike classical flow matching which regresses only a transport vector field, WFR-FM simultaneously regresses a vector field for displacement and a scalar growth rate function for birth-death dynamics, yielding continuous flows under the WFR geometry. Theoretically, we show that minimizing the WFR-FM loss exactly recovers WFR geodesics. Empirically, WFR-FM yields more accurate and robust trajectory inference in single-cell biology, reconstructing consistent dynamics with proliferation and apoptosis, estimating time-varying growth fields, and applying to generative dynamics under imbalanced data. It outperforms state-of-the-art baselines in efficiency, stability, and reconstruction accuracy. Overall, WFR-FM establishes a unified and efficient paradigm for learning dynamical systems from unbalanced snapshots, where not only states but also mass evolve over time. The Python code is available at https://github.com/QiangweiPeng/WFR-FM.

Streaming 4D Visual Geometry Transformer

cs.AI, q-bio.NC updates on arXiv.org

By: Dong Zhuo · Wenzhao Zheng · Jiahe Guo · Yuqi Wu · Jie Zhou · Jiwen Lu

1 April 2026 at 12:00

arXiv:2507.11539v2 Announce Type: replace-cross Abstract: Perceiving and reconstructing 3D geometry from videos is a fundamental yet challenging computer vision task. To facilitate interactive and low-latency applications, we propose a streaming visual geometry transformer that shares a similar philosophy with autoregressive large language models. We explore a simple and efficient design and employ a causal transformer architecture to process the input sequence in an online manner. We use temporal causal attention and cache the historical keys and values as implicit memory to enable efficient streaming long-term 3D reconstruction. This design can handle low-latency 3D reconstruction by incrementally integrating historical information while maintaining high-quality spatial consistency. For efficient training, we propose to distill knowledge from the dense bidirectional visual geometry grounded transformer (VGGT) to our causal model. For inference, our model supports the migration of optimized efficient attention operators (e.g., FlashAttention) from large language models. Extensive experiments on various 3D geometry perception benchmarks demonstrate that our model enhances inference speed in online scenarios while maintaining competitive performance, thereby facilitating scalable and interactive 3D vision systems. Code is available at: https://github.com/wzzheng/StreamVGGT.

A Multi-Modal CNN-LSTM Framework with Multi-Head Attention and Focal Loss for Real-Time Elderly Fall Detection

cs.AI, q-bio.NC updates on arXiv.org

By: Lijie Zhou · Luran Wang

25 March 2026 at 12:00

arXiv:2603.22313v1 Announce Type: cross Abstract: The increasing global aging population has intensified the demand for reliable health monitoring systems, particularly those capable of detecting critical events such as falls among elderly individuals. Traditional fall detection approaches relying on single-modality acceleration data suffer from high false alarm rates, while conventional machine learning methods require extensive hand-crafted feature engineering. This paper proposes a novel multi-modal deep learning framework, MultiModalFallDetector, designed for real-time elderly fall detection using wearable sensors. Our approach integrates multiple innovations: a multi-scale CNN-based feature extractor capturing motion dynamics at varying temporal resolutions; fusion of tri-axial accelerometer, gyroscope, and four-channel physiological signals; incorporation of a multi-head self-attention mechanism for dynamic temporal weighting; adoption of Focal Loss to mitigate severe class imbalance; introduction of an auxiliary activity classification task for regularization; and implementation of transfer learning from UCI HAR to SisFall dataset. Extensive experiments on the SisFall dataset, which includes real-world simulated fall trials from elderly participants (aged 60-85), demonstrate that our framework achieves an F1-score of 98. 7, Recall of 98. 9, and AUC-ROC of 99. 4, significantly outperforming baseline methods including traditional machine learning and standard deep learning approaches. The model maintains sub- 50ms inference latency on edge devices, confirming its suitability for real-time deployment in geriatric care settings.

Step 3.5 Flash: Open Frontier-Level Intelligence with 11B Active Parameters

cs.AI, q-bio.NC updates on arXiv.org

By: Ailin Huang · Ang Li · Aobo Kong · Bin Wang · Binxing Jiao · Bo Dong · Bojun Wang · Boyu Chen · Brian Li · Buyun Ma · Chang Su · Changxin Miao · Changyi Wan · Chao Lou · Chen Hu · Chen Xu · Chenfeng Yu · Chengting Feng · Chengyuan Yao · Chunrui Han · Dan Ma · Dapeng Shi · Daxin Jiang · Dehua Ma · Deshan Sun · Di Qi · Enle Liu · Fajie Zhang · Fanqi Wan · Guanzhe Huang · Gulin Yan · Guoliang Cao · Guopeng Li · Han Cheng · Hangyu Guo · Hanshan Zhang · Hao Nie · Haonan Jia · Haoran Lv · Hebin Zhou · Hekun Lv · Heng Wang · Heung-Yeung Shum · Hongbo Huang · Hongbo Peng · Hongyu Zhou · Hongyuan Wang · Houyong Chen · Huangxi Zhu · Huimin Wu · Huiyong Guo · Jia Wang · Jian Zhou · Jianjian Sun · Jiaoren Wu · Jiaran Zhang · Jiashu Lv · Jiashuo Liu · Jiayi Fu · Jiayu Liu · Jie Cheng · Jie Luo · Jie Yang · Jie Zhou · Jieyi Hou · Jing Bai · Jingcheng Hu · Jingjing Xie · Jingwei Wu · Jingyang Zhang · Jishi Zhou · Junfeng Liu · Junzhe Lin · Ka Man Lo · Kai Liang · Kaibo Liu · Kaijun Tan · Kaiwen Yan · Kaixiang Li · Kang An · Kangheng Lin · Lei Yang · Liang Lv · Liang Zhao · Liangyu Chen · Lieyu Shi · Liguo T

24 February 2026 at 13:00

arXiv:2602.10604v2 Announce Type: replace-cross Abstract: We introduce Step 3.5 Flash, a sparse Mixture-of-Experts (MoE) model that bridges frontier-level agentic intelligence and computational efficiency. We focus on what matters most when building agents: sharp reasoning and fast, reliable execution. Step 3.5 Flash pairs a 196B-parameter foundation with 11B active parameters for efficient inference. It is optimized with interleaved 3:1 sliding-window/full attention and Multi-Token Prediction (MTP-3) to reduce the latency and cost of multi-round agentic interactions. To reach frontier-level intelligence, we design a scalable reinforcement learning framework that combines verifiable signals with preference feedback, while remaining stable under large-scale off-policy training, enabling consistent self-improvement across mathematics, code, and tool use. Step 3.5 Flash demonstrates strong performance across agent, coding, and math tasks, achieving 85.4% on IMO-AnswerBench, 86.4% on LiveCodeBench-v6 (2024.08-2025.05), 88.2% on tau2-Bench, 69.0% on BrowseComp (with context management), and 51.0% on Terminal-Bench 2.0, comparable to frontier models such as GPT-5.2 xHigh and Gemini 3.0 Pro. By redefining the efficiency frontier, Step 3.5 Flash provides a high-density foundation for deploying sophisticated agents in real-world industrial environments.

Reading view