Today's Overview

Alibaba's Tongyi Qianwen Qwen3.7-Plus multimodal agent new foundation launched, ByteDance open-sourced Bernini unified AI video editing framework; Overseas, Microsoft Build 2026 focused on releasing the first reasoning model MAI-Thinking-1 and OpenClaw-derived assistant Scout, Anthropic officially filed IPO documents and deployed Claude Mythos to critical infrastructure in 15 countries.

Domestic Hot Topics

🔥 Alibaba Qwen3.7-Plus Launches, New Foundation for Multimodal Agents

Alibaba's Tongyi Qianwen Qwen3.7-Plus officially released, featuring multimodal agent capabilities, allowing one-click replication of desktop professional software workflows, further strengthening the Agent task execution foundation.

Source: [QbitAI] | Details: https://www.qbitai.com/2026/06/427730.html

🔥 ByteDance Open-Sources Bernini: Equipping DiT with an "LLM Strategist"

ByteDance open-sourced the unified framework Bernini, featuring "understand first, then act" AI video editing capabilities, using large language models as the decision-making hub to collaborate with diffusion models, enhancing controllability in video generation.

Source: [QbitAI] | Details: https://www.qbitai.com/2026/06/427810.html

🔥 Baidu Wenxin PaddleOCR-VL-1.6 Released, Accuracy Breaks 96.33%

Baidu Wenxin team released the PaddleOCR-VL-1.6 document parsing model, achieving 96.33% accuracy, setting a new SOTA for models of the same size, further consolidating its advantage in multimodal document understanding.

Source: [QbitAI] | Details: https://www.qbitai.com/2026/06/427754.html

🔥 Tsinghua AIR Open-Sources UniLab, Robot Training Enters "Minute-Level"

Tsinghua AIR open-sourced the UniLab humanoid robot motion control training framework, which can complete training on Mac and other edge devices in just 3 minutes, with speed improved 10x compared to before, lowering the barrier to embodied intelligence research and development.

Source: [QbitAI] | Details: https://www.qbitai.com/2026/06/427729.html

🔥 Stardust Intelligence Completes Over 1 Billion Yuan Series B Funding, Valuation Exceeds 10 Billion

Shenzhen embodied AI company Stardust Intelligence announces completion of over 1 billion yuan Series B funding round, valuation exceeding 10 billion yuan. This round of funding will be used for humanoid robot mass production and model iteration.

Source: [36Kr] | Details: https://36kr.com/p/3836068296209537

🔥 Zhejiang University Professor Team Secures Investment from Caitong and SenseTime, Entering High-Risk Scenario Embodied Robot Brain Market

Zhejiang University professor entrepreneurial team announces securing investment from institutions like Caitong and SenseTime, focusing on building embodied robot brains for high-risk operation scenarios, enhancing decision-making and execution capabilities in complex environments.

Source: [36Kr] | Details: https://36kr.com/p/3836744788014208

🔥 Doubao Launches Paid Services in Late June, Integrating with Douyin E-commerce

ByteDance's Doubao large model will officially launch paid subscriptions in late June and accelerate integration with the Douyin e-commerce ecosystem, marking a new phase in its commercialization path.

Source: [36Kr] | Details: https://36kr.com/p/3835814391100550

🔥 "Doubao Auto" Targets 100,000-200,000 Yuan Intelligent Vehicle Market

The "Doubao Auto" solution powered by the Doubao large model officially emerges, targeting the 100,000-200,000 yuan mainstream family car market segment, promoting large-scale deployment of large models in intelligent cockpit scenarios.

Source: [QbitAI] | Details: https://www.qbitai.com/2026/06/427760.html

International Hot Topics

🔥 Microsoft Build 2026: Launches First Reasoning Model MAI-Thinking-1

Microsoft unveils its first self-developed advanced reasoning model MAI-Thinking-1 at Build 2026, positioning it to compete with OpenAI's o-series and Claude's thinking mode, further improving the Azure AI ecosystem.

Source: [The Verge] | Details: https://www.theverge.com/tech/941664/microsoft-ai-model-reasoning-mai-thinking-1-build-2026

🔥 Anthropic Officially Files for IPO, Claude Mythos Deployed to Critical Infrastructure in 15+ Countries

Anthropic has officially filed its prospectus to冲刺上市 and is pushing Claude Mythos to critical infrastructure scenarios in more than 15 countries, further capturing the enterprise market.

Source: [TechCrunch] | Details: https://techcrunch.com/2026/06/02/anthropic-scales-claude-mythos-to-critical-infrastructure-in-15-countries/

🔥 Microsoft Releases Scout Personal Assistant and Project Solara AI Device System

Microsoft launched Scout, a personal assistant based on the OpenClaw concept, and simultaneously unveiled Project Solara—a new operating system specifically designed for AI Agent hardware devices.

Source: [The Verge] | Details: https://www.theverge.com/news/939713/microsoft-scout-assistant-openclaw

🔥 Anthropic Releases Cowork: A No-Code Claude Desktop Agent

Anthropic launched the Cowork product, bringing Claude desktop agent capabilities to ordinary users' local file workflows, enabling daily task automation without writing any code.

Source: [VentureBeat] | Details: https://venturebeat.com/technology/anthropic-launches-cowork-a-claude-desktop-agent-that-works-in-your-files-no

🔥 Google Pays Big Money to Play Developers to Purchase Code for Training AI

Google is quietly paying Play Store developers for source code to train its AI models, sparking a new round of discussions in the developer community about data sourcing and fairness.

Source: [Hacker News] | Details: https://www.404media.co/google-is-quietly-buying-code-from-play-store-developers-to-train-ai/

🔥 Google Redesigns Search Box – First Major Overhaul in 25 Years

Google has undertaken its first major redesign of the search box in 25 years, integrating Gemini AI capabilities and upgrading the search entry point from keyword-based retrieval to a multimodal task interface.

Source: [VentureBeat] | Details: https://venturebeat.com/technology/google-just-redesigned-the-search-box-for-the-first-time-in-25-years-heres-why-it-matters-more-than-you-think

🔥 Uber Burns Through Annual AI Budget in 4 Months, Emergency Cap on Employee Usage

Uber's internal AI computing expenditures exhausted its full-year budget in just 4 months, forcing the company to set caps on employee AI usage quotas, reflecting the widespread challenge of "budget overruns" in enterprise AI deployment.

Source: [TechCrunch] | Details: https://techcrunch.com/2026/06/02/uber-caps-employee-ai-spending-after-blowing-through-budget-in-four-months/

Further Reading