AI
Artificial intelligence: news, models, benchmarks and tools
DeepSeek V4.1-Flash-1017 Beats Claude Sonnet 5.6 on 10 of 11 Coding Tests While Using 20 Percent Fewer Tokens
DeepSeek V4.1-Flash-1017 reportedly beats Claude Sonnet 5.6 on 10 of 11 coding benchmarks with 20 percent fewer tokens. Here is...
Read moreDetails
ChatGPT Can Now Read Your Chrome Tabs: Tab Comparison, Work Mode, and the Permission Checks to Make First
OpenAI has turned its ChatGPT Chrome extension into a browser agent: tag Chrome in a conversation and it reads your...
Read moreDetails
Meta Just Released an 816-Billion-Parameter Model You Can Download. Here Is the Catch
Meta's Muse Spark Open is the biggest open-weight AI model release of 2026: 816 billion parameters, a million-token context window,...
Read moreDetails
Mistral’s ‘Le Chonk’ Is Europe’s Trillion-Parameter AI Bet: What Large 4 Means for Your API Bill
Mistral Large 4 is a 1.05-trillion-parameter open-weight model trained on 3,800 Nvidia GPUs in Europe. At .36/.18 per million tokens...
Read moreDetails
Apple Intelligence Is Finally Coming to China This December: iOS 27.3, Alibaba’s Models, and What Beijing Kept Out
After two years of regulatory delays, Apple Intelligence launches in China this December with iOS 27.3. Alibaba safeguards, government-approved model...
Read moreDetails
OpenAI Used Its Own AI to Catch an Alleged Spy in Its Boardroom
The FBI has arrested OpenAI engineer Abdulrahman Alfozan, 35, for allegedly planting a recording device disguised in toilet paper behind...
Read moreDetails
X’s New Subscription Puts Grok, Cursor, and Premium on One Bill: Every Tier From $8 to $200
SpaceX Super Intelligence has folded X Premium, Grok, and the Cursor coding assistant into one subscription with five tiers from...
Read moreDetails
Gemini 4 Argon Is Now Generally Available: $2/$10 Until November 30, Then the Price Doubles
Google's flagship Gemini 4 Argon model is now generally available to developers and subscribers worldwide. The $2/$10 introductory API price...
Read moreDetails
Microsoft’s Maia 300 Is Its Biggest Bet on Its Own Silicon: 288 GB of HBM4, 2nm Chiplets, and 40% More Tokens per Megawatt
Microsoft unveiled the Maia 300 on October 12, its third custom AI accelerator. With 288 GB of HBM4 memory, TSMC...
Read moreDetails
Meta’s First Frontier Model Is Here: Muse Wonder Costs $4 a Million Tokens, but the Best Version Stays Locked Away
Meta unveiled Muse Wonder and Muse Wonder Full on Monday, its first frontier AI models. The API price matches Claude...
Read moreDetails