AI Tools Nav
HomeToolsCompareGuideNewsSkillsOpen Source
中
AI Tools Nav

Curated AI tools directory — from choosing to mastering, all in one place.

RSSAPI

Navigation

  • Home
  • Tools
  • Compare
  • Guide
  • News
  • Skills
  • Open Source

Platform

  • Overview
  • API
  • RSS
  • Submit

About

  • About Us
  • Changelog
© 2026 AI Tools Nav - AI Tools Directory
AI News
94

AI News Brief | 2026-06-07

DeepSeek and Gemini tie for first in Shanghai Gaokao AI essay test, weekly roundups review a historic week in AI, OpenAI super app plans take shape, NVIDIA's Jensen Huang visits South Korea, AI coding tools shift from autocomplete to agent.

2026-06-07

AI Industry Overview

June 7 brought a moment of reflection after what multiple outlets described as "the most dramatic week in AI industry history." The defining story — Anthropic's "When AI Builds Itself" call for a global pause — continued to reverberate through policy circles and boardrooms, with a growing consensus that regardless of motivations, the disclosure of internal data has permanently elevated AI safety from a niche concern to a mainstream imperative.

On the product side, this week was a "model launch super-week": Microsoft's 7 MAI models, Google's Gemma 4 12B, NVIDIA's Nemotron 3 Ultra and RTX Spark, and MiniMax's M3 all debuted within days of each other. OpenAI's reported plans for the biggest ChatGPT overhaul since launch added another dimension to an already frenetic week. In China, DeepSeek V4 and Google Gemini 3.5 Flash tied for first place in a high-profile AI essay test using Shanghai's Gaokao (college entrance exam) prompt — a result that went viral on Chinese social media and demonstrated the competitive parity of Chinese LLMs in native-language creative tasks.


AI Takes the Gaokao: DeepSeek & Gemini Tie for First

The STAR Market Daily (科创板日报) tested six leading AI models on the 2026 Shanghai Gaokao essay prompt and published scored results:

Rank Model Score Grade
🥇 DeepSeek V4 66/100 Champion
🥇 Gemini 3.5 Flash 66/100 Champion
🥈 Kimi 2.6 (Moonshot AI) 63/100 A
🥉 Doubao 13.6 (ByteDance) 61/100 B
4 Qwen 3.7 (Alibaba) 58/100 B
5 GPT-5.5 (OpenAI) 53/100 C

The results, which quickly trended on Weibo, highlight the competitive strength of Chinese models in native-language creative writing. DeepSeek V4 tied with Gemini 3.5 Flash at the top, while Kimi 2.6 earned an A-grade. GPT-5.5's last-place finish was attributed to insufficient training depth on Chinese-language materials. The test was nicknamed "AI's Gaokao score release" by netizens.

Sina Finance / STAR Market Daily | Futunn


The Week in Review: June 1–7, 2026

Multiple outlets published comprehensive weekly roundups, calling this the most consequential week in AI history:

Anthropic's Brake-and-Accelerator Paradox: The company simultaneously warned of existential risk from recursive self-improvement (revealing 80% of its code is AI-generated) while racing toward a ~$965B IPO, launching Claude Opus 4.8, and expanding Mythos to 150+ organizations. The contradiction sparked accusations of regulatory capture but also forced a global conversation about institutional AI safety mechanisms.

Microsoft's Independence Declaration: The launch of 7 in-house MAI models at Build 2026 — spanning reasoning, coding, image, and speech — marked Microsoft's strategic pivot from OpenAI dependency to proprietary AI capability. Combined with Microsoft IQ, Scout, MDASH, and HorizonDB, the suite represents the most ambitious enterprise AI platform launch to date.

NVIDIA's Full-Stack Offensive: From the 550B Nemotron 3 Ultra orchestrator to the RTX Spark PC chip (with MediaTek) to the NemoClaw industrial agent platform, Jensen Huang laid out a vision spanning datacenter, personal computing, and industrial applications.

Open-Source Explosion: Google Gemma 4 12B (laptop-grade multimodal), NVIDIA Nemotron 3 Ultra (550B open-source MoE), LightAgent (agent orchestration framework), and Hermes Desktop (cross-platform AI agent) made this the richest week for open-source AI releases.

China's AI Ascent: Huawei Cloud united 20+ domestic model vendors, DeepSeek pushed forward with its first funding round, Alibaba's Qwen3.7-Plus beat Claude Opus 4.6 on Terminal-Bench 2.0, and DeepSeek tied Gemini on the Gaokao essay test — collectively signaling China's arrival as a full-spectrum AI power.

East Money Weekly | Sina Finance


NVIDIA's Jensen Huang Visits South Korea

NVIDIA CEO Jensen Huang began a 4-day visit to South Korea (June 5–8), meeting with executives from SK Group, Hyundai Motor, LG, NAVER, and gaming/robotics startups. He also appeared on tvN's hit variety show You Quiz on the Block and threw the opening pitch at a KBO baseball game. Korean media reported the visit will deepen collaboration on AI chips and memory semiconductors between NVIDIA and Korean firms.

Yonhap News


OpenAI Super App: More Details Emerge

Following the initial FT report, more details surfaced about ChatGPT's planned super app transformation. The redesigned app will reportedly organize around a three-layer architecture: a conversational foundation layer, a middle tools layer (coding, search, image generation), and a top services layer for third-party integrations. Early testers describe it as "Slack + Google Search + GitHub Copilot + App Store, all in one chat interface." Analysts view this as pre-IPO positioning — OpenAI needs to convince investors it's not just a model company, but a platform-level AI entry point.


Other Updates

  • Broadcom: Stock stabilized by week's end after the ~12.6% drop on AI chip guidance miss. Analysts remain bullish on long-term AI infrastructure demand.
  • Meta: Stock offering rumors persist; market focused on ROI for the $125–145B AI capex plan.
  • AI Coding Tools: GitHub Copilot, Cursor, v0, Claude Code, and Devin Desktop all shipped major updates this week. Developer consensus: the shift from autocomplete to autonomous agent is now irreversible.
  • ElevenLabs: Dubbing v2 and Stan Lee voice licensing continue to generate creative industry buzz.
  • Perplexity: CNN copyright lawsuit developments watched closely as a potential landmark case for AI training data boundaries.
  • Notion AI: Native integration of Claude Code, Cursor, and Codex as workspace agents signals a new paradigm of cross-tool AI collaboration.

This daily brief synthesizes 100+ sources into a coherent snapshot of the AI ecosystem as of June 7, 2026, and serves as a comprehensive weekly review for June 1–7.

Tools Mentioned

ChatGPT→Claude→Claude Code→Gemini→GitHub Copilot→DeepSeek→通义千问→Kimi→Midjourney→Cursor→v0 by Vercel→Perplexity→豆包→可灵→ElevenLabs→Pika→Sora→Devin→Windsurf→Notion AI→Aider→Cody→智谱清言→Bolt.new→海螺AI→秘塔 AI→
Featured
C
Freemium

ChatGPT

OpenAI's conversational AI assistant supporting text generation, coding, creative writing, and more.

ChatChatWritingCoding
Featured
C
Freemium

Claude

Anthropic's AI assistant, excels at long-text analysis, code review, and complex reasoning.

ChatChatCodingAnalysis
Featured
C
Freemium

Claude Code

Anthropic's terminal-native AI coding assistant with deep codebase understanding, multi-file editing, test generation, and Git integration.

AgentCodingTerminalEngineering
Featured
G
Freemium

Gemini

Google's multimodal AI model, deeply integrated with Google ecosystem, supporting text, image, and code understanding.

ChatChatMultimodalSearch
Featured
G
Freemium

GitHub Copilot

GitHub's AI coding assistant deeply integrated with VS Code, JetBrains, and other IDEs, supporting code completion and conversational coding.

AgentCode CompletionIDE IntegrationCoding
Featured
D
Freemium

DeepSeek

High-performance LLM from DeepSeek achieving GPT-4-level performance at a fraction of the cost, supporting 128K context and deep reasoning.

ChatChatCodingReasoning
Featured
Q
Freemium

Qwen

Alibaba's LLM series covering chat, coding, multimodal, and more, supporting long context and complex reasoning.

ChatChatCodingMultimodal
Featured
K
Freemium

Kimi

Moonshot AI's assistant known for ultra-long context (2M characters), excelling at document analysis, long-form summarization, and deep research.

ChatChatLong ContextAnalysis
Featured
M
Paid

Midjourney

Top-tier AI image generation tool, renowned for artistic style and high-quality output.

ImageImage GenArtDesign
Featured
C
Freemium

Cursor

VS Code-based AI-first code editor with powerful AI completions, inline editing, and codebase chat.

AgentEditorCodingCode Completion
Featured
v
Freemium

v0 by Vercel

Vercel's AI frontend generator that turns descriptions into React/Next.js UI components and page code.

AgentCodingFrontendUI Gen
Featured
P
Freemium

Perplexity

AI search engine combining LLM with real-time web search, providing accurate answers with citations and deep research mode.

SearchSearchResearchReal-time
Featured
D
Free

Doubao

ByteDance's AI assistant with text-to-image, voice chat, web search and other multimodal capabilities, excellent Chinese experience.

ChatChatMultimodalSearch
Featured
K
Freemium

Kling

Kuaishou's AI video generation tool, creating high-quality short videos from text and images with realistic physics effects.

VideoVideo GenShort VideoPhysics
Featured
E
Freemium

ElevenLabs

Leading AI voice synthesis platform supporting multilingual text-to-speech and voice cloning.

AudioTTSVoice CloneMultilingual
Featured
P
Freemium

Pika

AI video generation tool by Pika Labs, supporting text/image-to-video with rich creative effects and easy operation.

VideoVideo GenCreativeText-to-Video
Featured
S
Paid

Sora

OpenAI's text-to-video model, capable of generating high-quality videos up to one minute long.

VideoVideo GenText-to-Video
Featured
D
Paid

Devin

Cognition AI's fully autonomous AI software engineer that independently handles the full dev cycle from requirements to deployment.

AgentCodingAutomationFull Stack
Featured
W
Freemium

Windsurf

Codeium's AI IDE featuring the innovative Cascade flow agent mode, supporting auto-reasoning, multi-step editing and terminal operations.

AgentEditorCodingAutomation
Featured
N
Paid

Notion AI

Notion's built-in AI features for writing assistance, summarization, translation, and brainstorming.

OfficeWritingSummarizationCollaboration
A
Free

Aider

Terminal AI pair programming tool using Git for version control, supporting multiple LLM backends, excelling at multi-file refactoring and large codebases.

AgentCodingTerminalOpen Source
C
Freemium

Cody

Sourcegraph's AI coding assistant using code graph for full-repo context, supporting auto-fix, refactoring, and code generation.

AgentCodingCode CompletionContext
C
Freemium

ChatGLM

Zhipu AI's conversational model based on GLM architecture, supporting code generation, chart understanding, tool calling, and long text.

ChatChatCodingReasoning
Featured
B
Freemium

Bolt.new

StackBlitz's browser-based AI full-stack app generator that creates runnable web apps from prompts with one-click deploy.

AgentCodingFull StackDeploy
M
Freemium

MiniMax

MiniMax's multimodal AI assistant supporting text, image, audio, and video understanding and generation.

ChatChatMultimodalTTS
Featured
M
Free

MetaSo AI

AI search engine providing structured answers and mind maps, ideal for deep research and knowledge organization.

SearchSearchResearchMind Map

Related News

93

AI News Brief | 2026-06-06

OpenAI plans biggest ChatGPT overhaul into super app ahead of IPO, Microsoft Build products get first developer feedback, AI coding tools race intensifies, Anthropic debate pushes AI safety into mainstream, Qwen3.7-Plus impresses on benchmarks.

1d ago
Policy & Regs
94

AI News Brief | 2026-06-05

Anthropic pause controversy deepens as critics cry regulatory capture, Meta explores tens-of-billions stock offering for AI, Broadcom AI chip guidance misses, China rejects AI rivalry framing, Tencent says most code is now AI-generated.

2d ago
Policy & Regs
96

AI News Brief | 2026-06-04

Anthropic drops 'When AI Builds Itself' bombshell calling for global AI pause, Microsoft launches 7 MAI models at Build, Google open-sources Gemma 4 multimodal, NVIDIA unveils Nemotron 3 Ultra and RTX Spark chip, Huawei Cloud unites 20+ Chinese AI model vendors.

3d ago
← Back to AI News