open ai has a MAJOR problem with gpt 5.6 hallucination rates you see gpt 5.6 deleting entire codebases and deleting paid subscriptions in stripe accounts this is no joke how can a frontier model be trusted with a hallucination rate of 89% while an OPEN SOURCE model kimi k3 comes in at 49% ai labs in the US are in trouble chinese models have now caught up to frontier models and its not even close kimi k3 even beats fable 5 in most benchmarks

GPT-5.6 Sol Ultra crushed Fable 5 Max in the latest Lava Lamp test: better physics, better color, better graphics, and at 90% less token cost. If you're not counting the cost now, you better watch out when Fable 5 comes off your subscription in just two days. Anthropic better wake up here if they take it off your subscription. Most people will just jump straight to GPT-5.6.

this is crazy! this model is notably better than anything i have ever used even on /high effort is enough what are you guys using fable 5 for? you now have it through the 12th

get ready builders you will have access to the most capable model on the planet source: https://www.politico.com/news/2026/06/30/anthropic-wh-lifting-export-limits-00980865 that made up for the sonnet 5 FLOP

Only been able to make one of the “How to Run LLMs Locally” classes. Were the rest of the videos recorded?