CanItRun Logocanitrun.

Qwen3 8B

Qwen3 8B needs roughly 6.4 GB VRAM at Q4_K_M quantization (19.3 GB at FP16). 105 GPUs we track can run it fully in VRAM at 8k context.

105 GPUs run this natively · 2 with CPU offload

Alibaba8B params128k contextApache 2.0Commercial use ok

Qwen3 8B is a 8B parameter dense large language model developed by Alibaba. Released in April 2025, it is a text-only model with a 128K context window, released under the Apache 2.0 license, allowing commercial use. Supports thinking and non-thinking modes.

To run Qwen3 8B locally, you need approximately 6.4 GB of VRAM at Q4_K_M quantization with 8k context. 105 of the GPUs we track can run it fully in VRAM, with a further 2 able to offload to system RAM. At Q4_K_M it needs just 6.4 GB, making it accessible even on mid-range 16 GB cards like RTX 4080 and RTX 4070 Ti Super. At Q8_K_M (10.3 GB), you get near-FP16 quality while still fitting on 12, 16, 24, 48 and 80 GB GPUs. FP16 requires 19.3 GB, limiting it to 48 and 80 GB GPUs.

With an MMLU-Pro score of 56.73, it delivers strong general reasoning for local deployment. The license allows commercial use.

VRAM at each quantization

Assumes 8k context. KV cache grows linearly with context length.

QuantWeightsKV cacheTotal
FP3232.0 GB1.21 GB37.2 GB
BF1616.0 GB1.21 GB19.3 GB
FP1616.0 GB1.21 GB19.3 GB
Q8_08.0 GB1.21 GB10.3 GB
Q6_K6.6 GB1.21 GB8.7 GB
Q5_K_Mrec5.2 GB1.21 GB7.1 GB
Q4_K_M4.5 GB1.21 GB6.4 GB
Q3_K_M3.4 GB1.21 GB5.2 GB
Q2_K2.6 GB1.21 GB4.3 GB
NVFP4cuda4.0 GB1.21 GB5.8 GB

KV cache shown at 8k context (FP16). NVFP4 requires a CUDA GPU. Enable TurboQuant in the calculator to see reduced KV cache estimates.

Benchmarks

MMLU-Pro
56.73

GPUs that run Qwen3 8B natively (105)

Plus 2 GPUs that run it with CPU offload (slower)

Notes

Supports thinking and non-thinking modes.

Hugging Face ↗Ollama ↗Released 2025-04-29

Compare Qwen3 8B with other models

Frequently asked questions

What are the VRAM requirements for Qwen3 8B?
Qwen3 8B requires approximately 6.4 GB of VRAM at Q4_K_M quantization, 10.3 GB at Q8, and 19.3 GB at FP16. These numbers assume 8k context window; VRAM scales linearly with context length due to the KV cache.
How many parameters does Qwen3 8B have?
Qwen3 8B has 8 billion parameters.
How capable is Qwen3 8B?
With an MMLU-Pro score of 56.73, Qwen3 8B delivers solid general-purpose performance suitable for most everyday tasks and professional use.
Can Qwen3 8B run on a 16 GB GPU?
Yes. Qwen3 8B needs 6.4 GB at Q4_K_M, which fits in a 16 GB GPU like the RTX 4080 or RTX 4070 Ti Super.
What is the smallest quantization for Qwen3 8B that fits in 24 GB of VRAM?
At BF16, Qwen3 8B needs 19.3 GB — the highest-quality quantization that fits in 24 GB of VRAM.
What GPU do I need to run Qwen3 8B locally?
A 16 GB GPU is enough. At Q4_K_M, Qwen3 8B needs 6.4 GB VRAM. Good options: RTX 4080 (16 GB), RTX 4070 Ti Super (16 GB).