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) 90 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. Muse Spark 1.3 (max) 46 pt3. gpt-oss-120b (high) 39 pt4. GPT-5.6 Luna (max) 27 pt5. GPT-5.6 Terra (max) 22 pt

AI News

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

Grok Lite's Word Salad Error: What Grok.com Users Need to Know

Grok Lite's Word Salad Error: What Grok.com Users Need to Know

Grok Lite users on Grok.com experienced random word salad responses. xAI attributed the issue to a temporary generation glitch in the system

Micro1's 5x Revenue Jump Proves the Hunger for AI Training Data

Micro1's 5x Revenue Jump Proves the Hunger for AI Training Data

Micro1 grew its gross run rate from $100 million to $500 million in eight months. The startup is shifting toward synthetic data and off-the-

Slack Code Turns AI Coding Into a Multiplayer Experience

Slack Code Turns AI Coding Into a Multiplayer Experience

Slack launched Slack Code to bring AI coding agents into shared channels. The platform integrates agents like Devin and Claude Code for team

The Paradox of Anti-AI Fonts: Breaking Accessibility to Feed AI

The Paradox of Anti-AI Fonts: Breaking Accessibility to Feed AI

Anti-AI fonts designed to block scrapers inadvertently disable screen readers for disabled users. These obfuscation attempts serve as traini

Serval Catalyst Shifts IT Automation From Reactive To Proactive

Serval Catalyst Shifts IT Automation From Reactive To Proactive

Serval released Catalyst to automate enterprise IT workflows. The system uses a model-agnostic approach with GPT and Claude. Catalyst shifts

Pew Research Finds 35% of Post-ChatGPT Web Pages are AI-Generated

Pew Research Finds 35% of Post-ChatGPT Web Pages are AI-Generated

Pew Research found 35% of new web pages show AI authorship signs. AI content is 10 times more common on .com than .gov domains. Cloudflare r

"Don't Paste the AI, Please" Targets the LLM Copy-Paste Culture

"Don't Paste the AI, Please" Targets the LLM Copy-Paste Culture

A new satirical site called "Don't paste the AI, please" discourages LLM copy-pasting. The tool provides message links to remind colleagues

Why Huzzah Is Moving AI Coding From Chat Windows to .hz Files

Why Huzzah Is Moving AI Coding From Chat Windows to .hz Files

Huzzah introduces an experimental editor using pseudocode for AI prompts. The tool replaces transient chat messages with persistent .hz file

NanoClaw Deploys Full AI Agent Teams via a Single Slack Prompt

NanoClaw Deploys Full AI Agent Teams via a Single Slack Prompt

NanoCo launched NanoClaw to build AI agent teams in Slack via one prompt. The system uses Model Context Protocol to deploy persistent digita

Ramp Router: The Single API for Switching Between 8 AI Providers

Ramp Router: The Single API for Switching Between 8 AI Providers

Ramp launched Router to unify multiple AI models under one API. The service is free until 2026 and supports eight major AI providers. Ramp i

Meta AI Mac App Integrates Screen Awareness and Business Automation

Meta AI Mac App Integrates Screen Awareness and Business Automation

Meta AI launched a Mac app featuring system-wide dictation and screen awareness. The app integrates Meta Ads and Google Workspace to automat

Why Amazon Made Alexa+ Free for All Fire TV Users in the US

Why Amazon Made Alexa+ Free for All Fire TV Users in the US

Amazon removed the $19.99 monthly fee for Alexa+ on Fire TV devices. The AI assistant now offers conversational search and smart home integr

TerraPower's Molten Sodium Storage Solves AI Data Center Power Volatility

TerraPower's Molten Sodium Storage Solves AI Data Center Power Volatility

TerraPower will announce its first AI data center project this year. Meta has agreed to purchase eight Natrium plants for AI infrastructure.

Stripe's $7.5 Billion OpenRouter Deal Shifts Focus to AI Token Spend

Stripe's $7.5 Billion OpenRouter Deal Shifts Focus to AI Token Spend

Stripe acquired OpenRouter for 7.5 billion dollars to manage AI spend. The deal marks a shift from payment collection to token cost manageme

Why OpenAI's TAC Error Blocked Access to GPT-5.6 Sol

Why OpenAI's TAC Error Blocked Access to GPT-5.6 Sol

OpenAI revoked access to its Trusted Access for Cyber program for several researchers. The glitch affected Daybreak Blue users who utilize G

Why SpaceX Targeted Cognition Despite the CEO's Public Denial

Why SpaceX Targeted Cognition Despite the CEO's Public Denial

SpaceX reportedly attempted to acquire AI coding startup Cognition. Cognition CEO Scott Wu denied the sale but may discuss compute deals. Th

TrueForge Cuts AI Agent Costs by 75% Compared to Claude Managed Agents

TrueForge Cuts AI Agent Costs by 75% Compared to Claude Managed Agents

TrueFoundry released TrueForge as an open-source AI agent harness. The framework reduces task costs to $2.90 compared to $11.80 for managed

OpenAI Private Safety Processing Redefines Enterprise Data Sovereignty

OpenAI Private Safety Processing Redefines Enterprise Data Sovereignty

OpenAI launched Private Safety Processing to monitor AI misuse without storing data. The system detects malicious patterns across multiple s

How CUDA and OpenStreetMap Solved a Metadata-Free OSINT Challenge

How CUDA and OpenStreetMap Solved a Metadata-Free OSINT Challenge

A researcher used CUDA and OpenStreetMap to locate a resort from a single photo. The process filtered 80 million combinations using geometri

Google Gemini Shifts from Simple Search to 3D AI Learning Simulations

Google Gemini Shifts from Simple Search to 3D AI Learning Simulations

Google integrates 3D simulations and interactive visuals into Gemini. A new student hub allows for flashcards and custom practice quizzes. G

Why AI Companies Are Facing a Trust Crisis Despite Technical Gains

Why AI Companies Are Facing a Trust Crisis Despite Technical Gains

Public concern over AI has risen to 52 percent since 2021. AI firms are offering local bonuses to ease data center opposition. Industry lead

Qwen3.8-27B-Uncensored-GGUF Restores MTP for Faster Local Inference

Qwen3.8-27B-Uncensored-GGUF Restores MTP for Faster Local Inference

Qwen3.8-27B-Uncensored-GGUF is now available on Hugging Face. The model preserves MTP tensors to enable speculative decoding. Quantized vers

VentureBeat Recruits Rob Strechay to Lead Enterprise AI Research

VentureBeat Recruits Rob Strechay to Lead Enterprise AI Research

VentureBeat hired Rob Strechay as the first lead analyst for its research arm. A survey of 145 companies shows two-thirds adopt multi-model

Relativity Networks Secures $22M to Scale Hollow-Core Fiber for Data Centers

Relativity Networks Secures $22M to Scale Hollow-Core Fiber for Data Centers

Relativity Networks raised $22 million to deploy hollow-core fiber optics. The technology reduces data latency by 30% compared to traditiona

The $800 Billion Proposal to Move OpenAI Into Public Hands

The $800 Billion Proposal to Move OpenAI Into Public Hands

A proposal suggests Norway's GPF-G acquire OpenAI to end private AI monopolies. The plan argues that AI models are derivative works of the g

Snowflake Cuts AI Token Costs by 3x With Dynamic Model Routing

Snowflake Cuts AI Token Costs by 3x With Dynamic Model Routing

Snowflake introduced dynamic model routing to lower AI token costs. The system reduces expenses by up to 3x using an auto-routing option. Al

Claude Code Boosts Weekly Usage Limits by 50% for Paid Users

Claude Code Boosts Weekly Usage Limits by 50% for Paid Users

Anthropic increased Claude Code weekly limits by 50 percent. The promotion applies to Pro, Max, and Team plan users. Users can verify their

AI Agent Validation Gap: Why 85% of Failing Firms Automate Deployment

AI Agent Validation Gap: Why 85% of Failing Firms Automate Deployment

Nearly half of AI agents fail in production despite passing internal tests. Eighty-five percent of failing firms are moving toward automated

Why Agentic Commerce Fails Without a Unifying Execution Layer

Why Agentic Commerce Fails Without a Unifying Execution Layer

Commerce AI often fails due to a fragmented additive approach. A Unifying Execution Layer is required to maintain data consistency. Agentic

The 111% PR Surge That Redefined Linear's AI Workflow

The 111% PR Surge That Redefined Linear's AI Workflow

Linear saw a 111% increase in pull requests after AI adoption. AI agents are turning non-technical roles into active builders. Productivity