KO EN

AX BRIEF

AI news, benchmarks & engineering blog curation

● LIVE
지능1. Claude Fable 5.1 (max with fallback) 100 pt2. Claude Fable 5 (with fallback) 96 pt3. Claude Opus 5 (max) 95 pt4. GPT-6 Astra (max) 95 pt5. GPT-5.6 Sol (max) 91 pt코딩1. Claude Fable 5.1 (max) 100 pt2. Claude Opus 5 (xhigh) 94 pt3. GPT-6 Astra (max) 91 pt4. Muse Spark 1.3 (xhigh) 81 pt5. Grok 4.5 (high) 81 pt이미지1. GPT Image 2 (high) 100 pt2. MAI-Image-2.6 93 pt3. GPT Image 1.5 (high) 91 pt4. Reve 2.1 88 pt5. Muse Image 85 pt비디오1. Gemini Omni Flash 100 pt2. Minimax H3 Max (post-trained by fal) 94 pt3. Dreamina Seedance 2.0 720p 91 pt4. gemini-omni-1.1-flash 85 pt5. Wan 3.0 85 pt가격1. Claude Fable 5.1 (max with fallback) 100 pt2. GPT-6 Astra (max) 100 pt3. Claude Fable 5 (with fallback) 100 pt4. Claude Opus 5 (max) 49 pt5. GPT-5.6 Sol (max) 39 pt속도1. Gemini 3.5 Flash-Lite 100 pt2. Gemini 3.8 Flash (high) 85 pt3. Muse Spark 1.3 (max) 47 pt4. gpt-oss-120b (high) 43 pt5. GPT-5.6 Luna (max) 28 pt

AI News

Daily AI industry news — funding, products, policy, and major moves from global AI companies and startups.

Why Snap Spun Off Its AI Video Team into Dotmo

Why Snap Spun Off Its AI Video Team into Dotmo

Snap spun off its AI video team into a new company called Dotmo. CTO Bobby Murphy led the investment using his personal capital. The move sw

Why Elastic Paid $85 Million for DeductiveAI's AI SRE Tech

Why Elastic Paid $85 Million for DeductiveAI's AI SRE Tech

Elastic acquired DeductiveAI for up to 85 million dollars. The deal integrates AI-driven bug resolution into Elastic Observability. This mov

The $50 Billion Pivot: Why AWS May Sell Trainium Chips Externally

The $50 Billion Pivot: Why AWS May Sell Trainium Chips Externally

AWS is exploring the external sale of its Trainium AI chips. CEO Andy Jassy estimates a potential $50 billion annual revenue run rate. The m

Opus 4.5 Turns Application Code Into a Regeneratable Cache

Opus 4.5 Turns Application Code Into a Regeneratable Cache

Opus 4.5 reduces the cost of code production to near zero. Application code is shifting from a permanent asset to a regeneratable cache. Eng

Why SQLite Rejects External Contributions to Ensure Stability Until 2050

Why SQLite Rejects External Contributions to Ensure Stability Until 2050

SQLite ensures long-term stability by limiting external code contributions. The project uses aviation-grade DO-178B testing to eliminate bug

Why the US Government Blocked SK Telecom's Access to Claude Mythos

Why the US Government Blocked SK Telecom's Access to Claude Mythos

The US government revoked foreign access to Anthropic's Claude Mythos model. SK Telecom lost access due to historical business ties with Chi

Claude Code Artifacts Turns Terminal Sessions Into Live Webpages

Claude Code Artifacts Turns Terminal Sessions Into Live Webpages

Anthropic introduced Claude Code Artifacts for Team and Enterprise users. The tool converts terminal sessions into shareable 16 MiB HTML pag

Why OpenAI Recruited Noam Shazeer and Dean Ball Simultaneously

Why OpenAI Recruited Noam Shazeer and Dean Ball Simultaneously

OpenAI hired Transformer co-author Noam Shazeer and policy expert Dean Ball. The new Strategic Futures team manages regulatory risks and AI

Kē Turns Karamo Brown Into a Scalable AI Wellness Agent

Kē Turns Karamo Brown Into a Scalable AI Wellness Agent

Karamo Brown launched Kē, a wellness app featuring his AI digital clone. The app uses Delphi technology to provide voice-based personalized

Arbor Delivers 2.5x Performance Gains via Autonomous Optimization

Arbor Delivers 2.5x Performance Gains via Autonomous Optimization

Microsoft and Renmin University unveiled Arbor for autonomous AI optimization. The framework achieves 2.5x performance gains over standard c

Adobe Creative Agents Shift AI From Content Generation to Workflow Control

Adobe Creative Agents Shift AI From Content Generation to Workflow Control

Adobe introduced Creative Agents that control software APIs via natural language. These agents automate repetitive tasks across Premiere Pro

Why mdpresent Rejects LLMs for Deterministic Slide Rendering

Why mdpresent Rejects LLMs for Deterministic Slide Rendering

The mdpresent CLI tool converts Markdown into editable PPTX and PDF files. It uses a rule-based engine instead of LLMs to ensure determinist

Pixi Turns Static iMessage Stickers Into Interactive AI Characters

Pixi Turns Static iMessage Stickers Into Interactive AI Characters

Pixi launched an iOS app that sends interactive AI AR characters via iMessage. The service uses on-device AI to let characters react to the

AKB Merges Human Wikis and AI Context Into One Git Vault

AKB Merges Human Wikis and AI Context Into One Git Vault

Dnotitia released AKB as a shared knowledge base for humans and AI. The system uses Open Knowledge Format and MCP for seamless integration.

AWS Context Automates Knowledge Graphs to Power AI Agents

AWS Context Automates Knowledge Graphs to Power AI Agents

AWS launched AWS Context to automate knowledge graphs for AI agents. The system uses a feedback loop to learn from agent usage history. Inte

Anthropic Joins Frontier to Scale $1.8 Billion Carbon Removal

Anthropic Joins Frontier to Scale $1.8 Billion Carbon Removal

Anthropic became the first AI startup to join the Frontier carbon removal coalition. The coalition's total funding commitment has now reache

G7 Leaders Challenge US Control Over Anthropic AI Model Access

G7 Leaders Challenge US Control Over Anthropic AI Model Access

G7 leaders expressed concern over the US ability to revoke AI access. Anthropic's Mythos 5 and Fable 5 models are currently under export ban

The Qwen 3.6 27B Performance Gap That Redefines Local LLM Utility

The Qwen 3.6 27B Performance Gap That Redefines Local LLM Utility

Qwen 3.6 27B achieves a 77.2 score on the SWE-Bench Verified benchmark. Quantization on consumer GPUs introduces critical stability and hall

The Entity List Gap Leaving DeepSeek and CXMT Unchecked

The Entity List Gap Leaving DeepSeek and CXMT Unchecked

The US has paused adding over 100 Chinese firms to its trade blacklist. DeepSeek and CXMT remain active despite interagency approval for san

The 32% SaaS Crash: Why Proprietary Data Now Trumps Regulatory Moats

The 32% SaaS Crash: Why Proprietary Data Now Trumps Regulatory Moats

Public SaaS stocks saw a median valuation drop of 32 percent. AI agents are rendering seat-based pricing models obsolete. Proprietary domain

NVIDIA's ENPIRE: Robots that learn while we sleep

NVIDIA's ENPIRE: Robots that learn while we sleep

NVIDIA's GEAR Lab launched ENPIRE, a framework that lets AI agents train robots autonomously. It enables high-precision tasks like GPU insta

GLM-5.2 Hits 51 Intelligence Score to Lead Open-Weight Models

GLM-5.2 Hits 51 Intelligence Score to Lead Open-Weight Models

Z ai released GLM-5.2 as the top-performing open-weight model on the Intelligence Index. The model expands its context window to 1M tokens u

Meta's $14.8 Billion Scale AI Bet Turns 5,000 Engineers Into Labelers

Meta's $14.8 Billion Scale AI Bet Turns 5,000 Engineers Into Labelers

Meta invested $14.8 billion in Scale AI to prioritize data optimization. The company reassigned 5,000 engineers to perform manual data label

ktx Launches Self-Improving Context Layer for AI Data Agents

ktx Launches Self-Improving Context Layer for AI Data Agents

ktx released a self-improving context layer for AI data agents. The tool prevents aggregation errors like fan traps in data warehouses. It i

Google's Gemini Home Speaker Ends the Era of Rigid Voice Commands

Google's Gemini Home Speaker Ends the Era of Rigid Voice Commands

Google launched a $99.99 smart speaker powered by Gemini AI. A new premium subscription offers advanced Gemini Live capabilities. The device

The AI Moat: Why Human Relationships Outperform Transactional Efficiency

The AI Moat: Why Human Relationships Outperform Transactional Efficiency

AI is raising the productivity floor while commoditizing efficiency. Companies risk losing customer loyalty by over-automating relationships

Grok 4.1 Fast Outlasts Claude and GPT in 2D Battle Royale

Grok 4.1 Fast Outlasts Claude and GPT in 2D Battle Royale

Grok 4.1 Fast won a 2D battle royale against 10 other LLMs. The model achieved a 43 percent win rate with high cost efficiency. Over-alignme

Adam Collapses Complex CAD Menus Into a Single Text Prompt

Adam Collapses Complex CAD Menus Into a Single Text Prompt

Adam is an open-source web app that converts text into 3D CAD models. The tool generates parametric OpenSCAD files for easy modification. It

Pinterest Launches Ask Pinterest to Transform Shopping via Taste Graph AI

Pinterest Launches Ask Pinterest to Transform Shopping via Taste Graph AI

Pinterest has introduced Ask Pinterest as an experimental conversational AI app. The platform leverages its proprietary Taste Graph to provi

Claude Design Now Syncs with Claude Code to Enforce Brand Systems

Claude Design Now Syncs with Claude Code to Enforce Brand Systems

Anthropic updated Claude Design to support external design system imports. New bidirectional integration with Claude Code enables seamless s