@repligate @prerat Fable 5 is in a completely different league on postdoc-level math. historically Anthropic has massively underperformed on higher math (vs general intelligence), compared to OpenAI models which seem to have more math literature in the data mix. Fable 5 is the only exception so far