Edge Description: drop Input/Output bullet, add Benchmarks highlight

README.md

@@ -22,7 +22,7 @@ tags:
 Cosmos3-Edge is the most compact member of the Cosmos3 family — a 4B-parameter Omnimodal world model that generates coherent text, image, video, and action outputs from combinations of text, image, video, and action-trajectory inputs. It brings Cosmos3's multimodal understanding and world-generation capabilities to latency- and resource-constrained deployments.
 
 - **What's different from the larger models:** At 4B parameters, Cosmos3-Edge is the smallest model in the family — roughly 4× smaller than Cosmos3-Nano (16B) and 16× smaller than Cosmos3-Super (64B). It trades scale for efficiency so it can run under tighter compute, memory, and latency budgets, while keeping the same Mixture-of-Transformers (MoT) architecture and omnimodal interface as the larger models.
-- **Input / Output:** Text, image, video, and action-trajectory inputs in → text, image, video, and action outputs out. Video generation supports 256p and 480p.
+- **Benchmarks:** Cosmos3-Edge delivers the highest image-to-video generation throughput in the family while achieving competitive quality across PAIBench, RBench, and PhysicsIQ (evaluated at 480p, 24 fps on a single NVIDIA H100); see the [Benchmarks](https://huggingface.co/nvidia/Cosmos3-Edge#benchmarks) section.
 - **Example usage and output:** See the [Usage](https://huggingface.co/nvidia/Cosmos3-Edge#usage) section.
 - **Hardware:** The compact 4B size is intended for efficient single-GPU deployment; see [Usage](https://huggingface.co/nvidia/Cosmos3-Edge#usage) and [PBR](https://huggingface.co/nvidia/Cosmos3-Edge#pbr-performance-benchmark-reporting) for latency estimates.