Opus 5.5 is smarter, but Sol targets usability, poster argues
The author argues Anthropic optimizes for the smartest model while OpenAI's GPT-6 Sol targets the most-used one, near Fable 5.1 level at a tenth the cost and faster.
anthropic and openai have been trading the smartest-model title back and forth for a year opus 5.5 is smarter than gpt-6 sol but sol isn't trying to be the smartest. it's close, near fable 5.1 level, and it costs a tenth as much while running faster so anthropic optimized for being the best model. openai optimized for being the model people actually use not sure if openai's holding back something bigger internally to grab the top spot again later, or if they've just decided usability matters more than winning the intelligence race their next release will make that clear either way