Parameters
30B MoE · 3B activeOpen Model Catalogue
NVIDIA
Text generationReasoning
Nemotron 3.5 Lightning
nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA hybrid Mamba MoE — very fast on 1 × 80 GB
Min VRAM
≈ 78 GBHardware
1 × 80 GB GPUOpenAI-compatible API with a per-workload key — drop-in for existing SDKs.
Dedicated GPUs — your model, your instance, no shared queues.
Billed by the second with auto-shutdown — never pay for idle time you didn't plan.