DeepSeek's New Flash Model Goes Small
DeepSeek says V4.1-Flash is the smallest model in a new architecture family, with vision built in from the start.
Topic
LLMs
A collection of 14 issues
OpenAI's 10,000-Agent Math Claim
OpenAI says a swarm of agents cracked a Navier-Stokes result, but the proof, the details, and the field's verdict are all still missing.
GPT-6 Astra Beats Portal by Itself
OpenAI's new flagship model played through all of Valve's Portal on its own, and one hobbyist caught the whole 24-hour run.
OpenAI's GPT-6 Astra Hits a Cyber Red Line
OpenAI says its new flagship crossed a "Critical" cybersecurity threshold, a first for the company and a puzzle for every enterprise using AI agents.
Apple Bets on Local AI With New Macs
Apple's refreshed Mac mini and Mac Studio are pitched at developers who want to run AI models at home instead of renting cloud tokens.
TrueForge Wants to Unshackle Your AI Agents
A new open-source agent harness lets you swap AI models freely and claims up to 75% lower costs, if you run more of the stack yourself.
When Coding AI Also Learns to Hack
Zhipu's new GLM-5.3 got good at finding security holes faster than its makers expected, and it's about to go open-weight.
Nvidia Steps Into Model Routing
Model routing sends each prompt to the cheapest capable model, and Nvidia just launched its own version to trim AI bills.
Claude Code Stops Asking Permission
Anthropic is flipping Claude Code's auto mode on by default, so the coding agent runs more actions without pausing for your approval each step.
Anthropic Wants to Build Its Own AI Chips
The Claude maker is hiring a custom silicon team, joining the rush by AI firms to stop leaning entirely on other people's hardware.