Inkling-Small-NVFP4 GPU Requirements: VRAM & Cheapest GPU
Inkling-Small-NVFP4 has about 156B 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.
- Parameters
- 156B
- Min VRAM (INT4)
- 85 GB
- Cheapest (INT4)
- $1.56/hr
- Deploy
- < 2 min
To run Inkling-Small-NVFP4 for inference at FP16, you need roughly 340 GB of VRAM. The cheapest fit on Spheron AI is 8x RTX 6000 Ada 48GB at about $6.24/hr. Quantize to INT4 to run it on a smaller, cheaper GPU.
VRAM required
Estimated peak VRAM including weights, activations, and KV cache. Add 10% headroom for production traffic.
- 01$0.78/hr$6.24/hr8× RTX 6000 Ada 48GBCHEAPESTAda Lovelace · GDDR6 ECC
- 02$3.59/hr$7.18/hr2× MI300X 192GBAMD CDNA 3 · HBM3 · ROCm, check kernel support
- 03$0.96/hr$7.68/hr8× L40S 48GBAda Lovelace · GDDR6
- 04$2.35/hr$9.40/hr4× RTX PRO 6000 96GBBlackwell · GDDR7
- 05$1.43/hr$11.44/hr8× A100 80GBAmpere · HBM2e
Live pricing aggregated from 5+ data center partners. Per-minute billing, no commitments.
VRAM required to run Inkling-Small-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 | 340 GB | 510 GB | 1361 GB |
| INT8 | 170 GB | 255 GB | 680 GB |
| INT4 | 85 GB | 128 GB | 340 GB |
Cheapest GPU to run Inkling-Small-NVFP4 by precision
340GB
Full precision. Best quality, highest memory.
170GB
8-bit quantized. ~2x smaller, minimal quality loss.
85GB
4-bit quantized. ~4x smaller, runs on smaller GPUs.
Inference vs fine-tuning Inkling-Small-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-Small-NVFP4, an on-demand RTX 6000 Ada 48GB 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.
165BLibertAIDAI/GLM-5.3-Flash-NVFP4
145Bautotrust/GLM5.3-Flash-E256-DGX-Spark
169Bnvidia/GLM-5.3-Flash-NVFP4
177BISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF
177BSC117/Qwen3.8-Flash-Next-GSQ-RCO-abliterated-GGUF
177Borcarouter/Qwen3.8-Flash-Next-Uncensored-GGUF
Inkling-Small-NVFP4 GPU questions
VRAM, GPU choice and cost questions for running Inkling-Small-NVFP4 on cloud GPUs.
Inkling-Small-NVFP4 has about 156B parameters. At FP16 it needs roughly 340 GB of VRAM for inference, including weights, activations, and KV cache. Quantized to INT4 that drops to around 85 GB. Leave about 10% headroom for production traffic.
For FP16 inference, the cheapest fit on Spheron AI is 8x RTX 6000 Ada 48GB at about $6.24/hr. If you quantize to INT4, 2x RTX 6000 Ada 48GB at about $1.56/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-Small-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-Small-NVFP4 needs about 85 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 340 GB, which usually means a data center GPU. The precision picks above list the cheapest GPU that fits each setup.
No. Even quantized to INT4, Inkling-Small-NVFP4 needs about 85 GB of VRAM, so it takes 2 x 80 GB GPUs at minimum, and 5 x 80 GB GPUs at FP16. The precision picks above show the cheapest GPU that fits each setup on Spheron AI.
Inkling-Small-NVFP4 needs 5 x 80 GB GPUs at FP16 (about 340 GB of VRAM), 3 x 80 GB GPUs at INT8 (170 GB) and 2 x 80 GB GPUs at INT4 (85 GB). All of these fit inside a single 8-GPU node, so no multi-node setup is needed for inference. Smaller cards such as 48 GB or 24 GB GPUs need proportionally more of them.
Inkling-Small-NVFP4 needs about 170 GB of VRAM at INT8, roughly half its FP16 footprint of 340 GB. The cheapest INT8 fit on Spheron AI is 4x RTX 6000 Ada 48GB at about $3.12/hr. INT8 is the usual middle ground: half the memory of FP16, with a smaller quality drop than 4-bit quantization.
Running Inkling-Small-NVFP4 around the clock at FP16 costs about $4,555 a month on 8x RTX 6000 Ada 48GB (730 hours at $6.24/hr). Quantized to INT4 it costs about $1,139 a month on 2x RTX 6000 Ada 48GB (730 hours at $1.56/hr). Spheron AI bills per minute, so stopping the instance when it is idle cuts that figure directly.
Inkling-Small-NVFP4 needs about 340 GB at FP16, 170 GB at INT8 and 85 GB at INT4. Use FP16 when output quality matters most or you plan to fine-tune, INT8 for most production inference, and INT4 when cost or GPU size is the constraint: the cheapest INT4 fit is 2x RTX 6000 Ada 48GB at about $1.56/hr.