papers

Publications (22)

cs.CL2026

Kimi K2.5: Visual Agentic Intelligence

Kimi Team, Tongtong Bai, Yifan Bai +322

We introduce Kimi K2.5, an open-source multimodal agentic model designed to advance general agentic intelligence. K2.5 emphasizes the joint optimization of text and vision so that…

cs.AI2026

SkillRevise: Improving LLM-Authored Agent Skills via Trace-Conditioned Skill Revision

Yuxuan Liu, Zhaochen Su, Lingyun Xie +11

Agent skills are procedural artifacts that enable LLM agents to execute workflows, verify constraints, and recover from failures. Existing self-evolving methods refine skills using…

cs.CL2026

Evolutionary Guided Decoding: Iterative Value Refinement for LLMs

Zhenhua Liu, Lijun Li, Ruizhe Chen +5

While guided decoding, especially value-guided methods, has emerged as a cost-effective alternative for controlling language model outputs without re-training models, its effective…

cs.SE2026

Rethinking Self-Evolving Agent Skills: Feedback Dynamics over Multiple Rounds

Yuxuan Liu, Zhaochen Su, Yuhao Zhang +9

Self-evolving skill systems promise to improve agents by turning execution feedback into persistent skill updates without changing the underlying model. Yet it remains unclear when…

cs.AI2026

CostBench: Evaluating Multi-Turn Cost-Optimal Planning and Adaptation in Dynamic Environments for LLM Tool-Use Agents

Jiayu Liu, Cheng Qian, Zhaochen Su +4

Current evaluations of Large Language Model (LLM) agents primarily emphasize task completion, often overlooking resource efficiency and adaptability. This neglects a crucial capabi…

cs.CL2024

ConflictBank: A Benchmark for Evaluating the Influence of Knowledge Conflicts in LLM

Zhaochen Su, Jun Zhang, Xiaoye Qu +6

Large language models (LLMs) have achieved impressive advancements across numerous disciplines, yet the critical issue of knowledge conflicts, a major source of hallucinations, has…

cs.AI2026

The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence

MiniMax, :, Aili Chen +219

We introduce the MiniMax-M2 series, a family of Mixture-of-Experts language models built around the principle that mini activations can unleash maximum real-world intelligence. The…

cs.CL2022

Improving Temporal Generalization of Pre-trained Language Models with Lexical Semantic Change

Zhaochen Su, Zecheng Tang, Xinyan Guan +3

Recent research has revealed that neural language models at scale suffer from poor temporal generalization capability, i.e., the language model pre-trained on static data from past…

cs.AI2026

XSkill: Continual Learning from Experience and Skills in Multimodal Agents

Guanyu Jiang, Zhaochen Su, Xiaoye Qu +1

Multimodal agents can now tackle complex reasoning tasks with diverse tools, yet they still suffer from inefficient tool use and inflexible orchestration in open-ended settings. A…

cs.CL2024

Timo: Towards Better Temporal Reasoning for Language Models

Zhaochen Su, Jun Zhang, Tong Zhu +4

Reasoning about time is essential for Large Language Models (LLMs) to understand the world. Previous works focus on solving specific tasks, primarily on time-sensitive question ans…

cs.CV2025

OpenThinkIMG: Learning to Think with Images via Visual Tool Reinforcement Learning

Zhaochen Su, Linjie Li, Mingyang Song +8

While humans can flexibly leverage interactive visual cognition for complex problem-solving, enabling Large Vision-Language Models (LVLMs) to learn similarly adaptive behaviors wit…

cs.CL2026

Kimi K3: Open Frontier Intelligence

Kimi Team, Tongtong Bai, Yifan Bai +398

We introduce Kimi K3, a 2.8T parameter Mixture-of-Experts model with 104 billion activated parameters, native vision capabilities, and a 1-million-token context window. Kimi K3 is…

cs.AI2025

AdaCtrl: Towards Adaptive and Controllable Reasoning via Difficulty-Aware Budgeting

Shijue Huang, Hongru Wang, Wanjun Zhong +4

Modern large reasoning models demonstrate impressive problem-solving capabilities by employing sophisticated reasoning strategies. However, they often struggle to balance efficienc…

cs.LG2026

Advancing Multimodal Reasoning: From Optimized Cold Start to Staged Reinforcement Learning

Shuang Chen, Yue Guo, Zhaochen Su +7

Inspired by the remarkable reasoning capabilities of Deepseek-R1 in complex textual tasks, many works attempt to incentivize similar capabilities in Multimodal Large Language Model…

cs.CL2026

Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents

Shijue Huang, Hangyu Guo, Guanting Dong +8

Multimodal deep search requires an agent to solve open-world problems by chaining search, tool use, and visual reasoning over evolving textual and visual context. Two bottlenecks l…

cs.CV2024

SURf: Teaching Large Vision-Language Models to Selectively Utilize Retrieved Information

Jiashuo Sun, Jihai Zhang, Yucheng Zhou +3

Large Vision-Language Models (LVLMs) have become pivotal at the intersection of computer vision and natural language processing. However, the full potential of LVLMs Retrieval-Augm…

cs.CL2025

GRACE: Generative Representation Learning via Contrastive Policy Optimization

Jiashuo Sun, Shixuan Liu, Zhaochen Su +6

Prevailing methods for training Large Language Models (LLMs) as text encoders rely on contrastive losses that treat the model as a black box function, discarding its generative and…

cs.CL2026

The Tool Decathlon: Benchmarking Language Agents for Diverse, Realistic, and Long-Horizon Task Execution

Junlong Li, Wenshuo Zhao, Jian Zhao +18

Real-world language agents must handle complex, multi-step workflows across diverse Apps. For instance, an agent may manage emails by coordinating with calendars and file systems,…

cs.CV2025

Thinking with Images for Multimodal Reasoning: Foundations, Methods, and Future Frontiers

Zhaochen Su, Peng Xia, Hangyu Guo +12

Recent progress in multimodal reasoning has been significantly advanced by textual Chain-of-Thought (CoT), a paradigm where models conduct reasoning within language. This text-cent…

cs.CL2025

PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Models

Mingyang Song, Zhaochen Su, Xiaoye Qu +2

Process-level Reward Models (PRMs) are crucial for complex reasoning and decision-making tasks, where each intermediate step plays an important role in the reasoning process. Since…

cs.CV2026

AgentVista: Evaluating Multimodal Agents in Ultra-Challenging Realistic Visual Scenarios

Zhaochen Su, Jincheng Gao, Hangyu Guo +10

Real-world multimodal agents solve multi-step workflows grounded in visual evidence. For example, an agent can troubleshoot a device by linking a wiring photo to a schematic and va…

cs.CL2024

Living in the Moment: Can Large Language Models Grasp Co-Temporal Reasoning?

Zhaochen Su, Juntao Li, Jun Zhang +6

Temporal reasoning is fundamental for large language models (LLMs) to comprehend the world. Current temporal reasoning datasets are limited to questions about single or isolated ev…