NVIDIA Cosmos3-Super Models Hit Leaderboard: Text-to-Image Elo 1219, Pricing Coming Soon
ArtificialAnlys · x · 2026-07-31
Artificial Analysis updated its Text-to-Image and Image-to-Video leaderboards, adding NVIDIA's Cosmos3-Super-4Step models. On the text-to-image leaderboard, Cosmos3-Super-Text2Image (agentic) ranks 9-10 with an Elo of 1219, with pricing coming soon. The image-to-video model Cosmos3-Super-Image2Video-4Step joins the Video Arena. Top spots include OpenAI GPT Image 2 (Elo 1340), Reve 2.1 (Elo 1299), and Microsoft MAI-Image-2.5 (Elo 1270).
More from Multimodal
- Microsoft Open-Sources TRELLIS: Turns Any Image into 3D Assets in 3 Seconds — OwariDa · 2026-07-31
- Fish Audio S2.1 Pro Arrives on Melius, Benchmarked Against ElevenLabs v3 — yoobinray · 2026-07-31
- NVIDIA's Cosmos3 4-Step Distills Top Open-Weight Image-to-Video Leaderboards — ArtificialAnlys · 2026-07-31
- Klein Models Comparison for Image Reconstruction: Low-Res Photo Restoration Test — cradledust · 2026-07-31
- Turn iPhone Videos into Walkable 3D Models with Gauntlet Loops — mattshumer_ · 2026-07-31
- Google Earth Integrates Nano Banana for AI Image Generation — anselm · 2026-07-31