Santiago Valdarrama

@svpino

Another interesting area where PAN 2 really shines: multi-step plans. Remember that when working on a multi-step plan, if you pick the wrong step, the entire plan fails. PAN 2 got 75.1% of steps right and finished 34.0% of plans on a benchmark of 100 everyday tasks consisting of 3-6 steps. This was the highest full-plan completion among the well-known models IFM tested, such as Cosmos 3 Super and FLUX 3. More details: https://ifm.ai/pan/
打开原帖#511482
  1. Industry

    Elon Musk: @TheFlintFace @IterIntellectus 100GW in space is roughly equal to 800GW on Earth. Moreover, solar on Earth must be paired with massive batteries to continue powering co…
  2. Industry

    Runway: Seedance 2.5 in 4K is here
  3. Industry

    Perplexity: We are publishing new research on our findings from red-teaming SPACE…