NVIDIA’s Cosmos 3 Super claims 25x faster image and video generation
NVIDIAAI · x · 2026-07-22
NVIDIA says its new 4-step Cosmos 3 Super models can generate images and video up to 25x faster than the originals while still ranking near the top on open-weight benchmarks.
- The models are presented as open-weight releases.
- NVIDIA highlights #1 on image-to-video without audio and #2 on text-to-image on Artificial Analysis.
- The company says the models are available on Hugging Face.
More from Models
- No Covert Distillation: European AI Can't Beat Qwen2.5 3B Even With Logits — MaziyarPanahi · 2026-07-23
- BTL-3 launches as a 27B open-weight agent model in an 8.39GB file — QuixiAI · 2026-07-23
- Claude Opus 4.8 now drives 40% of Anthropic usage on OpenRouter — maferase · 2026-07-23
- GLM misses only “Computer locked” in nearly perfect ProgramBench cmatrix run — jyangballin · 2026-07-23
- GLM-5.2 hits 8.5% on ProgramBench, ranks 3rd overall — parth007_96 · 2026-07-23
- Gemini 3.5 Flash takes 14% of tokens but 40% of spend on OpenRouter — maferase · 2026-07-23