AI
Artificial intelligence: news, models, benchmarks and tools
What a Million Tokens Costs in October 2026: the Price List Every AI Builder Needs
From Haiku 5.5 at $0.10 to GPT-6 Astra at an estimated $10, we put every major AI model's October 2026...
Read moreDetails
Two DeepMind Researchers Quit Over Google’s Military AI Contracts, and They’re Naming Names
Sofia Liguori and Andreas Kirsch spoke on the record about why they left Google DeepMind: the Pentagons GenAI.mil deal, the...
Read moreDetails
OpenAI’s Models Are Coming to Google Cloud: What Changes When Rivals Share the Same Shelf
OpenAI is putting its frontier models and Codex agent workflows on Google Cloud's Model Garden, where enterprises can pay with...
Read moreDetails
Anthropic’s Claude Duet Is the First AI Assistant That Works Before You Ask
Anthropic's Claude Duet is a proactive AI agent app that briefs you, watches your connected apps, and handles tasks from...
Read moreDetails
China Just Ordered $35 Billion of Nvidia GPUs. Nvidia Says It Can Only Deliver a Fraction
ByteDance, Alibaba and Tencent have ordered more than 3 million RTX PRO 5500 workstation GPUs at $11,400 each, but Nvidia...
Read moreDetails
Europe’s Biggest Independent Cloud Is Training Its Own AI Model: OVHcloud’s Early-2027 Bet Against Mistral and the EU Gigafactories
OVHcloud CEO Octave Klaba told Reuters the cloud operator expects to launch its first proprietary AI model in early 2027,...
Read moreDetails
DHS Ordered US Agencies to Purge DeepSeek AI Three Months Ago, Nobody Was Told
US ordered federal civilian agencies to purge DeepSeek AI tech under a July order revealed only in October, with vague...
Read moreDetails
An AI Helped Jailbreak the iPhone 18 Pro, and That Changes the Math for Everyone
Calif.io demonstrated an untethered jailbreak of an iPhone 18 Pro on iOS 27.0.1 at Hexacon, and says Anthropic's Claude Mythos...
Read moreDetails
Nvidia’s New Open Models Run AI Agents in Your Data Center and on Your PC
Nvidia's Nemotron 3.5 family puts a 120-billion-parameter agent model in enterprise data centers and smaller Bolt models on AI PCs,...
Read moreDetails
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