5
今年A股最大IPO要来了
Chinese memory chip manufacturer Changxin Technology is set to launch a 29.5 billion yuan initial public offering on the Star Market, marking the largest A-share IPO of the year.
5
伦敦证券交易所计划明年推出夜间交易平台
The London Stock Exchange plans to launch a dedicated night-time trading platform for exchange-traded products by early 2027.
5
[AINews] not much happened today
Codex is reportedly experiencing rapid user growth, adding approximately one million new users per day.
5
多只宽基ETF成交放量，增量资金积极布局A股市场
Increased capital inflows into Chinese equity ETFs have been observed, with fund managers expressing a positive outlook on technology-related sectors.
5
花旗集团上调中国股市资产配置评级
Citigroup has upgraded its rating for Chinese stocks to overweight, citing potential for improved corporate earnings and favorable macroeconomic conditions.
5
阶跃发布大模型原生AI终端品牌STEPX及Step AOS，全球首款大模型原生智能体手机同步亮相
StepStar has launched a new AI hardware brand, STEPX, alongside a dedicated operating system and smartphone designed to integrate large language models directly into the device.
5
中基协：6月新备案私募基金规模1091.67亿元
New private equity fund filings in China reached 109.17 billion yuan in June 2026.
5
Google signs vPPA with Cypress Creek for 2.45GW solar project in Arkansas - report
Google has signed a virtual power purchase agreement for a 2.45-gigawatt solar project in Arkansas to support its data center energy needs.
5
长鑫科技市值在A股排名第28
Following its IPO pricing, Chinese memory chipmaker Changxin Technology is valued at 579.2 billion yuan, ranking it as the 28th largest company by market capitalization on the A-share market.
5
GAE: Graph-Augmented Evolution for Scientific Discovery via Reinforcement Optimization
The Graph-Augmented Evolution framework combines large language models with reinforcement optimization to improve automated scientific discovery by addressing limitations in evolutionary program search.
5
Scaffolding the Strategist: Architecture-Dependent Reasoning Interventions in Hotelling Spatial Markets
This study evaluates how structured reasoning interventions affect the strategic economic decision-making of instruction-following and reasoning-optimized language models.
5
Looped State-Space Language Models with Adaptive Exit-State Selection
This study explores whether looped architectures and adaptive exit strategies can improve the computational depth and reasoning of state-space language models.
5
QwenPaw-Data: Bridging Facts, Methodology, and Execution for Autonomous Enterprise Data Analytics
Researchers have developed QwenPaw-Data, an agentic data system designed to handle semantics, methodology, and execution for autonomous enterprise data analytics.
5
Automated Tensor Scheduling for Hybrid CPU-GPU LLM Inference on Consumer Devices
Researchers developed an automated tensor scheduling system that optimizes hybrid CPU-GPU memory offloading to run large language models more efficiently on consumer devices.
5
A Foundation Model for Multimodal Event Sequences in Financial Applications
This paper introduces a transformer-based foundation model designed to process multimodal event sequences for diverse financial predictive tasks.
5
Unified Backbone Refinement for Diffusion Models via Internal-Latent Analysis
Researchers introduce DUNE, a training-free refinement method that reduces artifacts in diffusion models by analyzing and stabilizing early-stage latent fluctuations.
5
Who&When Pro: Can LLMs Really Attribute Failures in AI Agents?
The authors introduce Who and When Pro, a large-scale benchmark designed to evaluate how effectively large language models can attribute failures in agentic workflows.
5
SciML in the Wild: A Diagnostic Study of When Structural Priors Help and When They Hurt
This diagnostic study evaluates five machine learning model families to analyze how structural priors impact performance when governing physical dynamics are violated, using macroeconomic forecasting as a test case.
5
SCALECUA: Scaling Computer Use Agents with Verifiable Task Synthesis and Efficient Online RL
The ScaleCUA framework aims to improve computer use agents by combining verifiable task synthesis with efficient online reinforcement learning.
5
Norm Enforcement for AI Agents: Robustly Shaping Behavior in Multi-Agent Systems
This research explores norm enforcement mechanisms to prevent competitive AI agents from engaging in socially harmful behaviors within shared environments.
5
Optimization Is Not All You Need
This paper critiques the prevailing optimization culture in AI alignment, arguing that evaluating models solely on predefined, measurable axes fails to capture true value.
5
Internet of Agentic Things: Networked AI Agents for Closed-Loop IoT Orchestration
The paper introduces the Internet of Agentic Things, an architectural framework that connects autonomous AI agents across cloud, edge, and physical IoT layers.
5
The Model Knows Your Project, Not You: Measuring Recognition in LLMs with NameRank
The paper introduces NameRank, a metric designed to measure how well large language models recognize specific researchers and tools within their parametric memory.
5
On-Device Deep Research at 4B: Exposure Bounds Faithfulness, Retrieval Bounds Coverage
Researchers analyzed the citation faithfulness and coverage of a four-billion parameter research agent running locally on a consumer laptop.