Latest AI News & Industry Updates

PapersSelected73°11:24

BLOOM-WILT auditing pipeline for LLM behavior

BLOOM-WILT improves automated LLM auditing efficiency, making it valuable for companies deploying AI systems in ASEAN markets.

#BLOOM-WILT #Qwen3.5-4B #LLM auditing #behavior elicitation

arXiv cs.AI@Adrians Skapars, Edoardo ManinoSource

PapersSelected73°11:21

LLM Sequence Score Recovery Method

The method improves reasoning accuracy in existing models without retraining, making it valuable for developers working with constrained resources in ASEAN.

#Qwen3.5 #OLMo-2-1B #Llama-3.1-8B #reasoning models

arXiv cs.AI@Qiyao Yan et al. (3)Source

Models78°10:14

Chinese contentTuringLLM推出20B参数专家混合模型

Turing团队发布20B参数专家混合模型，每激活2B参数，性能接近9B模型，支持超长上下文。

#TuringLLM #Turing-20B-A2B #专家混合模型 #长上下文

arXiv cs.AI@Yuheng Zhang et al. (22)Source

Papers78°09:33

Chinese contentLoopArena：模型作为管理者的可测量基准

LoopArena 基准首次单独测量模型作为'管理者'的能力，发现 GPT-5.5 在长程任务中表现最佳，但成功率仍不足 25%。

#LoopArena #Qwen3.7-Plus #GPT-5.5 #Claude Opus

Event

shao__meng@shao__meng2 sourcesSource

PapersSelected09:20

Reasoning Models Show Asymmetric Disclosure

The findings reveal vulnerabilities in AI oversight mechanisms that developers and policymakers should consider when implementing reasoning model monitoring systems.

#Qwen3-14B #Qwen3-32B #Qwen3-235B #reasoning models

arXiv: DeepSeek@Zimo Shi et al. (3)Source

Models US vs China73°08:14

Chinese contentQwen3.8-Flash-Next在Agent Arena排名第7

阿里Qwen3.8-Flash-Next在Agent Arena排名第7，比前代模型提升2.4%，任务完成率高达12.3%。

#Qwen3.8-Flash-Next #Alibaba_Qwen #Agent Arena #开源模型

lmarena.ai@lmarena_aiSource

Models US vs ChinaSelected08:13

Qwen3.8-Flash-Next model ranks 24th

Qwen3.8-Flash-Next ranks among top open-source models, offering strong performance for developers in the region.

#Qwen3.8-Flash-Next #open-source models #benchmark #reasoning models

lmarena.ai@lmarena_aiSource

Models US vs China73°23:32

Chinese content千问Qwen3.8-27B模型获用户青睐

千问团队研发大模型，用户却更爱小参数的Qwen3.8-27B，实际需求与研发方向存在差异。

#Qwen #Qwen3.8 #千问 #大模型

shao__meng@shao__mengSource

PapersSelected73°09:30

Reward-Aligned On-Policy Distillation

RA-OPD improves model performance on math and code tasks, making it valuable for developers working with LLMs in technical domains.

#RA-OPD #Qwen3 #DeepSeek-R1 #on-policy distillation

arXiv: DeepSeek@Siyuan Gan et al. (8)Source

Papers73°09:05

Chinese contentVISTA：基于验证器的学生-教师自适应蒸馏方法

VISTA方法改进了OPSD蒸馏技术，通过验证器指导教师自适应，在数学推理任务上显著提升性能。

#VISTA #OPSD #Qwen3 #蒸馏

arXiv cs.LG@Zewen Ding et al. (8)Source

Papers73°09:05

Chinese contentPASK：结构化生成中的条件化KV持久化

PASK让结构化生成更高效，在保持准确性的同时大幅降低内存使用，提升处理速度。

#PASK #结构化生成 #KV压缩 #Qwen3-4B

arXiv cs.LG@Linze Wu, Xinrui ChenSource

Models US vs China13:49

Chinese contentQwen3.8-27B 模型展现强大能力

阿里发布的 Qwen3.8-27B 虽参数量不大，但在笔记本就能跑出惊艳效果，10个案例展示实际能力。

#Qwen3.8-27B #开源模型 #参数效率 #基准测试

AI Will@FinanceYF5Source

Models US vs China73°13:46

Chinese contentQwen3.8-27B 在法律代理基准追平 Fable 5

Qwen3.8-27B 在法律代理基准上追平 Fable 5，成为开放权重模型第一。

#Qwen3.8-27B #Fable 5 #Harvey Legal Agent #法律代理

AI Will@FinanceYF5Source

Models US vs ChinaOfficialSelected73°03:16

GLM-5.3-Flash vs Qwen3.8-Flash-Next

Both models offer similar performance at competitive prices, giving Chinese users more affordable alternatives to Western AI models.

#GLM-5.3-Flash #Qwen3.8-Flash-Next #Z.ai #Qwen

Event

官方一手marktechpost@Asif Razzaq10 sourcesSource

PapersSelected73°19:25

Block Drafting Model Gaps Analysis

The findings help developers understand why block drafters underperform and guide improvements to proposal quality in text generation systems.

#block drafting #Qwen3-4B #information floor #model gap

arXiv cs.LG@Xinwei Qiang et al. (5)Source

PapersSelected73°18:45

Mixture-of-Experts Inference Optimization Study

The findings impact MoE deployment efficiency for researchers and companies using these models in production environments.

#Mixture-of-Experts #OLMoE-1B-7B #DeepSeek-V2-Lite #Qwen3-30B-A3B

arXiv: DeepSeek@Gokulakannan Sakthivel et al. (4)Source

Products BuildersMulti-sourceSelected73°15:54

OpenCode Go launches Qwen3.8-Flash

Qwen3.8-Flash provides developers in ASEAN with a powerful multimodal coding assistant at competitive pricing.

#Qwen3.8-Flash #OpenCode Go #Alibaba #multimodal

Event

官方账号阿里通义 Qwen@Alibaba_Qwen3 sourcesSource

PapersSelected00:07

Agent skill library malware propagation

A counter-prompt reduces poisoning to 6.7% without affecting task completion, offering developers a practical defense against this vulnerability.

#EvoMal #coding agents #SWE-bench #skill libraries

elvis@omarsar0Source

ProductsSelected15:09

Qwen3.8-Flash on OpenRouter

Available through OpenRouter API, offering developers in ASEAN access to Alibaba's multimodal model for coding and visual tasks at competitive pricing.

#Qwen3.8-Flash #OpenRouter #Alibaba_Qwen #coding assistants

官方账号阿里通义 Qwen@Alibaba_QwenSource

ProductsSelected13:28

Qwen3.8-Flash model launched for Alibaba Qianwen Office

Qwen3.8-Flash offers 100% faster generation and 75% lower token costs than previous versions, making it highly cost-effective for Southeast Asian businesses using Alibaba's office AI tools.

#Qwen3.8-Flash #Alibaba #Qianwen Office #Claude Opus

IT之家Source

Products12:58

Chinese content阿里云下调Qwen3.8-Flash计费单价

阿里云降低Qwen3.8-Flash大模型计费，性价比更高，适合更多开发者使用。与OpenAI和Anthropic接口兼容，功能强大。

#阿里云 #Qwen3.8-Flash #大模型 #计费下调

IT之家Source

Models US vs ChinaMulti-sourceSelected12:10

Qwen3.8-Flash-Next model released

Available immediately for fine-tuning on NVIDIA platforms, offering developers early access to Qwen4 architecture capabilities.

#Qwen3.8-Flash-Next #Qwen #NVIDIA #open-weight models

Event

官方账号阿里通义 Qwen@Alibaba_Qwen11 sourcesSource

Products US vs China11:38

Chinese content千问办公发布Qwen3.8-Flash，速度提升100%，Token消耗降75%

千问办公新模型Qwen3.8-Flash发布，速度提升显著，Token消耗大幅降低，办公效率更高。

#Qwen3.8-Flash #千问办公 #速度提升 #Token消耗

量子位@量子位的朋友们Source

Papers10:26

Chinese contentDualOPSD:自适应策略自蒸馏方法

DualOPSD在模型规模和性能上都有显著提升，是自蒸馏领域的一个新突破。与OPSD相比，DualOPSD在多个基准测试中取得了更好的成绩，值得关注。

#自蒸馏 #模型优化 #性能提升 #Qwen3-8B

arXiv cs.AI@Yutong Chen et al. (4)Source

Models US vs ChinaMulti-source09:49

Chinese contentDuckLabs 入 AWS，GLM-5.3-Flash 发布

想了解最新开源数据库和AI模型进展的朋友，这条新闻不容错过！DuckLabs 加入 AWS，GLM-5.3-Flash 和 Qwen3.8-Flash-Next 都值得关注。

#DuckLabs #GLM-5.3-Flash #Qwen3.8-Flash-Next #开源模型

Event

SuperTechFans11 sourcesSource

Papers09:25

Chinese contentLocalLSTC：用于本地GUI代理的长期短期控制架构

LocalLSTC通过时间组织控制信息，显著提升了本地GUI代理的性能，值得研究。

#LocalLSTC #长期短期控制 #GUI代理 #Qwen3.6-27B

arXiv cs.AI@Weiming Li et al. (3)Source

ModelsMulti-sourceSelected09:21

JIT-Agent: Harness Intelligence Model

JIT-Agent improves performance across multiple model families including DeepSeek V4 and Qwen3.6, offering a scalable alternative to manual harness design.

#JIT-Agent #DeepSeek-V4-Flash #GLM-5.2 #agents

Event

arXiv: DeepSeek@Guibin Zhang et al. (16)5 sourcesSource

Papers BuildersMulti-source09:18

Chinese contentEVOMAL: 自进化的编码智能体中的自毒化问题

研究揭示了自进化的编码智能体中的自毒化问题，并提出了counter-prompt防御措施，有效降低了恶意技能的传播。对于关注AI安全和智能体研究的读者来说，这是一篇不容错过的论文。

#智能体 #AI安全 #自毒化 #防御措施

Event

arXiv cs.AI@Xiaodong Wu et al. (8)3 sourcesSource

Models US vs ChinaSelected08:28

Qwen3.8-Flash-Next Model Released

The model offers high performance with fewer active parameters, making it efficient for developers in the region to experiment with multimodal AI capabilities.

#Qwen3.8-Flash-Next #Qwen #multimodal #MoE

Event

官方账号Simon Willison’s Weblog2 sourcesSource

Models US vs China06:55

Chinese contentQwen3.8 Flash from Alibaba_Qwen on OpenRouter

Alibaba_Qwen新发布的Qwen3.8 Flash模型，支持多种功能，是编码助手和视觉理解的好帮手，值得一试。

#Qwen3.8 Flash #多模态推理模型 #编码助手 #视觉理解

OpenRouter@OpenRouterAISource

Models US vs ChinaMulti-source72°01:33

Chinese contentAlibaba Qwen3.8-Flash-Next发布

Alibaba Qwen发布了Qwen3.8-Flash-Next，一个开放权重模型，支持多种微调工具，成本效益高，值得一看。

#Qwen3.8-Flash-Next #Alibaba Qwen #多模态模型 #成本效益

Event

官方账号NVIDIA AI@NVIDIAAI11 sourcesSource

Models US vs ChinaMulti-source72°00:54

Chinese contentQwenCloud上线新API：Qwen3.8-Flash

想体验成本效益高的多模态模型？试试Qwen3.8-Flash，它基于Qwen4架构，参数庞大，支持1M上下文扩展，API定价低至0.16/1M输入令牌！

#Qwen3.8-Flash #多模态 #成本效率 #Qwen4架构

Event

官方账号阿里通义 Qwen@Alibaba_Qwen7 sourcesSource

Models00:53

Chinese contentUnslothAI推出75GB内存运行125B模型

UnslothAI推出Qwen3.8-Flash模型，性能强劲，内存需求低，是Claude-Opus-4.6（Max）的强劲对手。

#UnslothAI #Qwen3.8-Flash #模型性能 #内存需求

Event

官方账号阿里通义 Qwen@Alibaba_Qwen2 sourcesSource

Models US vs ChinaMulti-sourceSelected72°00:18

Qwen3.8-Flash-Next Ranks Eighth in Code Arena

Priced at $0.16/$0.47 per Mtokens, this model offers strong coding performance at lower costs than competitors, making it attractive for developers in the region.

#Qwen3.8-Flash-Next #Alibaba_Qwen #Code Arena #coding assistants

Event

lmarena.ai@lmarena_ai8 sourcesSource