Matt Wolfe

@mreflow

People are pointing out that it depends on effort level. Lower effort levels are more token efficient... However, it seems a bit misleading when the benchmarks they are showing on their site are based on Max effort level.
打开原帖#511482
  1. Frontier

    Florida asks a court to halt OpenAI
  2. Frontier

    Sonnet 5.5 is faster and still trails Opus
  3. Frontier

    Nvidia puts the switch on the model path