Inkling-NVFP4 GPU Requirements: VRAM & Cheapest GPU
Inkling-NVFP4 has about 553B parameters. See exactly how much GPU memory it needs at FP16, INT8, and INT4, and the cheapest GPU to run it, with live hourly pricing from 5+ data center partners.
To run Inkling-NVFP4 for inference at FP16, you need roughly 1205 GB of VRAM. The cheapest fit on Spheron is 8x B200 192GB at about $42.72/hr. Quantize to INT4 to run it on a smaller, cheaper GPU.
Estimated peak VRAM including weights, activations, and KV cache. Add 10% headroom for production traffic.
- 01$5.34/hr$42.72/hr8× B200 192GBCHEAPESTBlackwell · HBM3e
- 02$5.81/hr$46.48/hr8× B300 288GBBlackwell Ultra · HBM3e
Live pricing aggregated from 5+ data center partners. Per-minute billing, no commitments.
VRAM required to run Inkling-NVFP4
Estimated peak VRAM at context length 4,096 and batch size 1, including weights, activations, and KV cache. Quantizing to INT8 (Q8) or INT4 (Q4) cuts memory roughly in half and in quarter.
| Precision | Inference | LoRA fine-tune | Full fine-tune |
|---|---|---|---|
| FP16 | 1205 GB | 1808 GB | 4821 GB |
| INT8 | 603 GB | 904 GB | 2410 GB |
| INT4 | 301 GB | 452 GB | 1205 GB |
Cheapest GPU to run Inkling-NVFP4 by precision
Full precision. Best quality, highest memory.
8-bit quantized. ~2x smaller, minimal quality loss.
4-bit quantized. ~4x smaller, runs on smaller GPUs.
Inference vs fine-tuning Inkling-NVFP4
Inference only holds the model weights plus a KV cache, so it is the cheapest setup. LoRA fine-tuning adds a small adapter and roughly 50% more memory. Full fine-tuning holds gradients and optimizer state on top of the weights, which is about 4x the inference footprint, so it often needs multiple GPUs even when inference fits on one. For Inkling-NVFP4, an on-demand B200 192GB instance covers inference and LoRA, while a full fine-tune needs several times that memory and often spans multiple GPUs. Check the live GPU pricing for current rates.
Similar models
Compare GPU requirements for models in the same class.
Inkling-NVFP4 GPU questions
Inkling-NVFP4 has about 553B parameters. At FP16 it needs roughly 1205 GB of VRAM for inference, including weights, activations, and KV cache. Quantized to INT4 that drops to around 301 GB. Leave about 10% headroom for production traffic.
For FP16 inference, the cheapest fit on Spheron is 8x B200 192GB at about $42.72/hr. If you quantize to INT4, 4x A100 80GB at about $3.40/hr runs it for less. Pricing is aggregated live from 5+ data center partners with per-minute billing.
LoRA fine-tuning adds roughly 50% on top of inference memory, so it usually fits the same class of GPU. Full fine-tuning holds gradients and optimizer state and needs about 4x the inference VRAM, which often means multiple GPUs for Inkling-NVFP4. The VRAM matrix above shows the exact estimate for each setup.
We read the parameter count directly from the model's safetensors metadata on HuggingFace, then estimate peak VRAM from weights, activations, KV cache, and framework overhead at your chosen precision. The estimate lands within about 15% of real-world use for most transformer models.
Quantized to INT4, Inkling-NVFP4 needs about 301 GB of VRAM, so it is too large for a single 24 GB RTX 4090 and needs a bigger card or multiple GPUs. At FP16 it needs roughly 1205 GB, which usually means a data center GPU. The precision picks above list the cheapest GPU that fits each setup.