NVIDIA · Vision · 120B parameters · 12B active · OpenMDW-1.1
NVIDIA Nemotron 3.5 Super VL 120B-A12B
Needs 144 GB at FP8 / INT8. The smallest real configuration today is 8× Tesla V100 32GB at Vast.ai, $0.54 per hour, about $391 a month at 730 hours.
Parameters120B12B active per token
Context1M tokensmaximum published
Checkpoint249 GBBF16, as published
LicenseOpenMDW-1.1check before commercial use
ReleasedOct 15, 2026documented date
Smallest configuration · FP8 / INT8
8× Tesla V100 32GB256 GB VRAM · 56% used
- Per hour
- $0.54
- Per month
- $391
- Cheapest at
Vast.ai
What is documented about NVIDIA Nemotron 3.5 Super VL 120B-A12B
news checked 2026-10-27NVIDIA's new flagship open model: a vision-language model for reasoning, coding, tool use, documents and video. Card: '120B total parameters and 12B active', up to 1M context, text, image and video input. 88 blocks = 40 Mamba-2 + 40 latent-MoE + 8 attention, so only the 8 attention layers hold a KV cache. 512 experts, 22 active. HF counts 124.45B. Card 'Release Date: Hugging Face 10/15/2026'. A Base-BF16 checkpoint is also published.
- https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16
- https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16/raw/main/config.json
- https://huggingface.co/api/models/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16?blobs=true
Memory · 3 precisionsHow much VRAM it needs
Weights plus runtime overhead. The context dimension appears once the model's architecture is on file.
| Precision | Weights | Weights + headroom |
|---|---|---|
| BF16published checkpoint | 249 GB | 299 GB |
| FP8 / INT8estimated | 120 GB | 144 GB |
| INT4 / MXFP4estimated | 72 GB | 87 GB |
Cross-checked on Hugging Face Oct 28, 2026: 124.5B parameters in the safetensors headers, 248.9 GB of weight files. Method: weights as published at the native precision, verified on Hugging Face 2026-10-28; other precisions estimated at 1 (8-bit) and 0.6 (4-bit) bytes per parameter; × 1.2 runtime headroom; KV cache not added (architecture not readable).
Smallest configuration per GPU · 110 GPUs · cheapest eligible on-demand offer per GPU count, every provider on fileWhat it costs to run, GPU by GPU
The fewest GPUs of one type whose total memory fits the model at short context, priced from the cheapest eligible on-demand offer for exactly that count (45-day window, newest reading per provider), with the next providers beside it. Monthly = 730 hours.
| GPU | Smallest config | VRAM used | $ / hour | $ / month | Cheapest at |
|---|---|---|---|---|---|
| Tesla V100 32GBBest fit | 8× 32 GB256 GB total | 56% | $0.54 | $391 | Vast.aiTrust 78 · read Nov 1 · receipt · also Clore.ai $0.69, Lium $4.40 |
| A800 80GB PCIe | 2× 80 GB160 GB total | 90% | $0.96 | $702 | Vast.aiTrust 78 · read Nov 1 · receipt |
| Quadro RTX 8000 | 4× 48 GB192 GB total | 75% | $1.01 | $741 | Vast.aiTrust 78 · read Nov 1 · receipt |
| RTX 3090 | 8× 24 GB192 GB total | 75% | $1.17 | $851 |  Clore.aiTrust 46 · read Nov 1 · receipt · also Vast.ai $1.29, Lium $2.32 |
| A100 80GB | 2× 80 GB160 GB total | 90% | $1.20 | $878 | Vast.aiTrust 78 · read Nov 1 · receipt · also Thunder Compute $2.18, Spheron $2.38 |
| L40S | 4× 48 GB192 GB total | 75% | $1.52 | $1,110 |  LiumTrust 40 · read Nov 1 · receipt · also Vast.ai $1.87, VoltageGPU $3.04 |
| A100 40GB | 4× 40 GB160 GB total | 90% | $1.60 | $1,170 | Vast.aiTrust 78 · read Nov 1 · receipt · also iwinv $3.66, AtmosCompute $4.05 |
| MI300X | 1× 192 GB192 GB total | 75% | $1.65 | $1,205 |  GPUBankTrust 93 · read Oct 25 · receipt · also OpenRelay $1.99, Runpod $2.39 |
| RTX PRO 6000 | 2× 96 GB192 GB total | 75% | $1.67 | $1,218 | Vast.aiTrust 78 · read Nov 1 · receipt · also UpCloud $1.94, Edgevana $3.07 |
| RTX A6000 | 4× 48 GB192 GB total | 75% | $1.68 | $1,226 | LiumTrust 40 · read Nov 1 · receipt · also Spheron $1.77, GPU.ai $1.88 |
| RTX PRO 5000 | 4× 48 GB192 GB total | 75% | $1.92 | $1,403 | Vast.aiTrust 78 · read Nov 1 · receipt · also Edgevana $4.60, AtmosCompute $5.28 |
| RTX PRO 4000 | 8× 24 GB192 GB total | 75% | $1.93 | $1,406 | Vast.aiTrust 78 · read Nov 1 · receipt · also Edgevana $2.15, AtmosCompute $3.47 |
| Gaudi 2 | 2× 96 GB192 GB total | 75% | $2.11 | $1,538 |  Cyfuture AInot scored · read Nov 1 · receipt |
| RTX 4000 Ada | 8× 20 GB160 GB total | 90% | $2.14 | $1,559 | Vast.aiTrust 78 · read Nov 1 · receipt |
| MI325X | 1× 256 GB256 GB total | 56% | $2.25 | $1,643 |  OpenRelaynot scored · read Nov 1 · receipt · also Cyfuture AI $3.14, DigitalOcean $3.80 |
| RTX 6000 Ada | 4× 48 GB192 GB total | 75% | $2.40 | $1,753 | Vast.aiTrust 78 · read Nov 1 · receipt · also GPU.ai $2.60, io.net $2.73 |
| A100X | 2× 80 GB160 GB total | 90% | $2.56 | $1,870 | Vast.aiTrust 78 · read Nov 1 · receipt |
| A40 | 4× 48 GB192 GB total | 75% | $2.60 | $1,898 |  Denvr CloudTrust 67 · read Nov 1 · receipt · also Exoscale $4.18 |
| L40 | 4× 48 GB192 GB total | 75% | $2.65 | $1,932 |  io.netTrust 42 · read Nov 1 · receipt · also GPU.ai $3.12, QuantaCloud $3.75 |
| RTX 4090 | 8× 24 GB192 GB total | 75% | $2.99 | $2,184 | Vast.aiTrust 78 · read Nov 1 · receipt · also Sesterce Cloud $3.52, Spheron $3.84 |
| RTX 5880 Ada | 4× 48 GB192 GB total | 75% | $3.20 | $2,339 | Vast.aiTrust 78 · read Nov 1 · receipt · also AtmosCompute $5.77 |
| H100 | 2× 80 GB160 GB total | 90% | $3.42 | $2,496 |  GcoreTrust 80 · read Nov 1 · receipt · also GPU.ai $3.96, UpCloud $4.00 |
| B200 | 1× 180 GB180 GB total | 80% | $3.75 | $2,738 |  packet.ainot scored · read Nov 1 · receipt · also Beam $4.11, UpCloud $4.11 |
| MI350X | 1× 288 GB288 GB total | 50% | $3.99 | $2,913 | OpenRelaynot scored · read Nov 1 · receipt |
| The 24 cheapest of 39 GPU types that fit. Every GPU, any precision and context, in the finder | |||||
NVIDIA Nemotron 3.5 Super VL 120B-A12B · related