Why Aren't People Talking About This Google Model?
Matt Wolfe showed 5 AI tools in this video on . None are verified available in the EU yet. Each row below jumps to the exact second. Part of week 2026-36.
Tools shown in this video, with timestamps
Post-trained update of Fable 5 with fewer cybersecurity refusals, anti-distillation measures and unchanged $10/$50 per million token pricing. Tops Artificial Analysis (66 vs Opus 5's 63) but is the most expensive per task; the creator's Mega Bonk clone was excellent but cost about $114 in usage.
Live demoReleased alongside Fable 5.1 but only available to a small group of cyber defender companies.
Cheap, fast coding-strong model scoring about 74% on DeepSWE, tied with Claude Opus 5 and above Fable 5, at $2.36 per task versus much higher for those. Creator considers it underhyped; it ranks only around 10th on Artificial Analysis overall and uses many output tokens.
Live demoUpcoming OpenAI model announced as delayed for safety; it can find and exploit security vulnerabilities, reaching roughly 40% success versus GPT-5.6 Soul's 11.5% with fewer tokens. Release expected soon.
Training technique reportedly used for Astra that processes the same text multiple times, obscuring chain-of-thought reasoning. This raises concerns about monitoring and safety.
