NVIDIA models
Which NVIDIA models can you run on Modal?
NVIDIA
Nemotron 3.5 Super VL 120B A12B NVFP4
NVIDIA’s multimodal model for visual reasoning and agents: 120B total parameters, 12B active, up to 1M context, and NVFP4 weights.
Text VisionNVIDIA
Nemotron 3.5 Lightning 30B A3B NVFP4
NVIDIA's fast open agent workhorse: a hybrid Mamba-2 MoE with 30B total and 3B active parameters, 1M-token context, built-in speculative decoding, and OpenMDW-1.1 weights.
TextNVIDIA
NVIDIA Nemotron 3 Ultra 550B A55B NVFP4
NVIDIA's frontier-scale open reasoning model: a 550B latent MoE with 55B active parameters, up to 1M-token context, NVFP4 weights, and OpenMDW-1.1 licensing.
TextNVIDIA
NVIDIA Nemotron 3 Super 120B A12B NVFP4
NVIDIA's hybrid Mamba-Transformer MoE for agentic reasoning: 120B total and 12B active parameters, up to 1M-token context, native NVFP4 pretraining, and multi-token prediction.
Text
| Model | Architecture | Context | Modalities | Input / Output |
|---|---|---|---|---|
| Nemotron 3.5 Super VL 120B A12B NVFP4 | 120B MoE · 12B active | Up to 1M | Text · Image | Dedicated only |
| Nemotron 3.5 Lightning 30B A3B NVFP4 | 30B MoE · 3B active | 1M | Text | Dedicated only |
| NVIDIA Nemotron 3 Ultra 550B A55B NVFP4 | 550B MoE · 55B active | 1M | Text | Dedicated only |
| NVIDIA Nemotron 3 Super 120B A12B NVFP4 | 120B MoE · 12B active | 1M | Text | Dedicated only |
Dedicated Endpoint compute is billed by the second. See compute pricing.
Latest updates

Scaling Intelligence: A landmark commission and exhibition

Sandbox resourcing: burst, don't reserve
Why you're paying for capacity you aren't using, and how to pay only for what you use.

Modal becomes a gVisor maintainer
Modal runs upwards of 40 million gVisor containers every day. Here's why we bet on it, what we've built into it, and where we want to take it next.
