Compare LLM Models
Select 2-5 models to compare side by side.
Select models to compare
0 / 5 selected
Top downloads
Select 2-5 models
| Attribute |
nvidia
|
|---|---|
| Company | nvidia |
| Release Date | 2026-05-26 |
| Parameters (B) | 3 |
| Architecture | Transformer |
| Context Window | 24000 |
| Input Modalities | text, image |
| Open Source | Yes |
| License | other |
| Score | 0 |
| VRAM (GB) | 12 |
| Compute | NVIDIA GPU with CUDA (Ampere A100, Hopper H100, Blackwell, or Lovelace RTX 4090/L40), Linux, BF16 inference; ~12 GB VRAM with la_flash sparse attention at 4K, up to ~35 GB with dense SDPA on high-resolution images |
| Pricing (Input) | — |
| Pricing (Output) | — |