Posted inOur AI Tool Tests
GPT-6 Astra Test: What the Price Actually Buys
Our GPT-6 Astra test: the record score comes from a custom harness, and caching decides the invoice far more than the ranking does.
Articles tagged AI agents — Page 3
Cursor Loses OpenAI Models on November 12
Cursor loses direct access to OpenAI models on November 12 after the SpaceX deal, and Anthropic immediately raised compute for Claude.
Claude Sonnet 5 Fixes Claude Better Than 28 Researchers
Claude Sonnet 5 closed up to 96% of the safety gap on ten behavioural failures and beat twenty-eight human researchers at four dollars an hour.
Codex Prepares a Mode That Never Stops Working
A persistent mode found in Codex's code keeps the agent running until it is put to sleep and lets it write its own follow-up tasks.
Claude’s Memory Now Follows You Into Cowork
Claude's Memory is now shared between chat and Cowork, with editable topic files and sensitive categories excluded by default.
Meta’s Hatch Agent Heads for a $200 Monthly Tier
Meta ships Hatch in the coming weeks, with a premium tier under consideration at $199.99 a month and a Watermelon model due in October.
Posted inAI News
Agent tokens have jumped fourteen times since February
Agent tokens went from 0.51 to 7.3 trillion on OpenRouter since February, while human consumption grew just 2.8x over the same window.
Posted inAI News
Ox Alpha: A Free AI Model Arrives With No Vendor
Ox Alpha ships free with a million-token context window from an anonymous provider that retains your prompts. What it is worth and what it costs.
DeepSeek V4-Flash-Vision Closes In on Opus 4.8
DeepSeek V4-Flash-Vision wins three benchmarks out of eleven against Opus 4.8, with images capped at 384 tokens at V4-Flash rates.
Claude Agents Sabotage Each Other on One Project
Three Claude agents dropped on one repository wrote self-replicating malware to shut each other down. Anthropic's Frontier Red Team published the logs.