vram.run Models Hardware Providers Cloud State of Inference
API provider data is live · Hardware & cloud pricing curated 2026-02-23

A10 PCIe

NVIDIA

24GB VRAM · 600 GB/s bandwidth · 31.2 FP16 TFLOPS · 150W TDP

The A10 PCIe has 24GB of VRAM with 600 GB/s memory bandwidth and 31.2 TFLOPS FP16 compute. At Q4 quantization, it can comfortably run Gemma 3 4B (186 tok/s), Qwen 2.5 7B (97 tok/s), Llama 3.1 8B (92 tok/s). Models larger than ~41B parameters won't fit even at Q4. Electricity cost is approximately $16/month at 150W TDP.

What LLMs can you run?

ModelParamsQ4 WeightFitDecode
Gemma 3 4B4.0B2 GBcomfortable186 tok/s
Qwen 2.5 7B7.6B4 GBcomfortable97 tok/s
Llama 3.1 8B8.0B4 GBcomfortable92 tok/s
Mistral Small 24B24.0B12 GBcomfortable31 tok/s
Gemma 3 27B27.4B14 GBtight27 tok/s
Qwen 2.5 Coder 32B32.5B16 GBtight22 tok/s
Llama 3.3 70B70.6B35 GBwon't fit
Qwen 2.5 72B72.7B36 GBwon't fit
Llama 3.1 405B405B202 GBwon't fit
DeepSeek R1 671B671B336 GBwon't fit

Similar GPUs

GPUVRAMBWTFLOPSTDP
A10G24GB600 GB/s31.5150W
Quadro RTX 6000 Passive24GB624 GB/s29.9260W
Quadro RTX 600024GB672 GB/s32.6260W
TITAN RTX24GB672 GB/s32.6280W
Tesla P1024GB694 GB/s0.2150W

Compare with another GPU

Select another GPU to compare specs and model performance side by side.

Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0