Local LLM model fit

Can my GPU run Qwen3 8B?

Qwen3 8B is a 8B Qwen model. This page estimates Q4 VRAM fit, Ollama command, context planning, and fallback choices for common local AI GPUs.

Check Qwen3 8B in the calculator

Q4 runtime estimate6 GB
Ollama commandollama run qwen3:8b
Recommended GPURTX 4060 8GB or better

Best use

Fast general local assistant with reasoning/coding balance. Weakness: Less capable than 14B/32B models for large tasks.

GPU fit table

HardwareExamplesClean capacityQ4 needStatusCalculator
6 GB VRAM entry GPUGTX 1660, RTX 2060 6GB4.5 GB clean VRAM6 GBRAM offloadOpen calculator
8 GB VRAM mainstream GPURTX 3060 Ti, RTX 4060, RTX 30706.5 GB clean VRAM6 GBRuns locallyOpen calculator
10 GB VRAM older high-end GPURTX 3080 10GB8.5 GB clean VRAM6 GBRuns locallyOpen calculator
12 GB VRAM local agent GPURTX 3060 12GB, RTX 4070, RTX 507010.5 GB clean VRAM6 GBRuns locallyOpen calculator
16 GB VRAM creator GPURTX 4060 Ti 16GB, RTX 4080, RTX 5070 Ti, RTX 508014.5 GB clean VRAM6 GBRuns locallyOpen calculator
24 GB VRAM homelab workstationRTX 3090, RTX 409022.5 GB clean VRAM6 GBRuns locallyOpen calculator
32 GB VRAM Blackwell workstationRTX 509030.5 GB clean VRAM6 GBRuns locallyOpen calculator
48 GB VRAM workstationRTX A6000, L40S 48GB46.5 GB clean VRAM6 GBRuns locallyOpen calculator
Apple Silicon 32 GB unified memoryM2 Max 32GB, M3 Max 36GB26 GB unified6 GBRuns locallyOpen calculator
Apple Silicon 256 GB unified memoryMac Studio M3 Ultra 256GB, Mac Studio M4 Ultra 256GB250 GB unified6 GBRuns locallyOpen calculator

Quantization memory estimate on a 12GB GPU preset

QuantizationEstimated memoryUse case
Q4 / 4-bit6 GBDefault local inference balance
Q5 / 5-bit7.5 GBBetter quality, more VRAM
Q8 / 8-bit12 GBHigh quality, much more VRAM
FP16 / 16-bit24 GBMostly workstation/server use

Data sources and confidence

This is a practical planning estimate, not a benchmark. Real memory use changes with backend, context length, KV cache, quantization file, drivers, and offloading settings.

Verified

2026-05-19

Confidence

medium