GPT-6 Sol called a "delusional disaster" as user backs Opus 5.5 as new SOTA
The post calls Sol a delusional disaster, says Luna is only decent for simple tasks, and argues Opus 5.5 is the new SOTA while criticizing OpenAI's marketing of smaller models.
GPT 6 Sol is a delusional disaster. GPT 6 Luna is decent for simple tasks. Opus 5.5 seems to be the new SOTA, it genuinely surprised me with its efficiency and effectiveness. For a long run, I only used around 4–6%. We’re reaching a point where people start asking: Why do they release garbage models like Sol and Luna? "They are smaller models, idiot!!" The usage rate of these smaller models shouldn’t be hyped as "WOW, THE REVOLUTION". Because then they should expect backlash if they perform poorly compared to the competition. Even Astra wasn’t an insane leap forward in quality, yet it was marketed as the new AGI SOTA... this kind of marketing is embarrassing. Less talking, more results, OpenAI. At the moment, Anthropic has delivered a strong release. Hopefully it doesn’t get quantized, and therefore lobotomized, in a day.