RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can I run Qwen3 4B Instruct 2507 locally?

Qwen3 4B Instruct 2507 by Qwen has 4.0B parameters. At 4-bit with an 8K context it needs about 4.1 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track. See all models that fit in 8 GB of VRAM.

4.1 GBQ4_K_M · 8K context
RTX 4060smallest consumer GPU that fits
M1 16GBsmallest Mac that fits
256Kmax context
GQAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from unsloth/Qwen3-4B-Instruct-2507-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K) •rough2.8 GB3.4 GB6.7 GB20.2 GB
3-bit (Q3_K_M) •usable3.2 GB3.7 GB7.1 GB20.6 GB
4-bit (Q4_K_M) •good, most popular3.6 GB4.1 GB7.5 GB21.0 GB
5-bit (Q5_K_M) •very good3.9 GB4.5 GB7.9 GB21.4 GB
6-bit (Q6_K) •near lossless4.3 GB4.9 GB8.3 GB21.8 GB
8-bit (Q8_0) •effectively lossless5.2 GB5.8 GB9.2 GB22.7 GB
16-bit (F16/BF16) •original weights8.8 GB9.3 GB12.7 GB26.2 GB

Which GPUs and Macs run Qwen3 4B Instruct 2507?

Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).

HardwareMemoryFits?Best quantNeedsMax tok/s (est.)
RTX 3060 12GB12 GBFitsF169.3 GB45details
RTX 4060 Ti 16GB16 GBFitsF169.3 GB36details
RTX 309024 GBFitsF169.3 GB116details
RTX 409024 GBFitsF169.3 GB125details
RTX 509032 GBFitsF169.3 GB223details
RTX 5070 Ti16 GBFitsF169.3 GB111details
RX 7900 XTX24 GBFitsF169.3 GB119details
Mac M4 Pro 48GB48 GBFitsF169.3 GB34details
Mac M4 Max 128GB128 GBFitsF169.3 GB68details
Mac M3 Ultra 512GB512 GBFitsF169.3 GB102details
DGX Spark128 GBFitsF169.3 GB34details
RTX 407012 GBFitsF169.3 GB63details
RTX 5060 Ti 16GB16 GBFitsF169.3 GB56details
RTX 40608 GBFitsQ8_05.8 GB64details
Arc B58012 GBFitsF169.3 GB57details
RTX 3060 Ti8 GBFitsQ8_05.8 GB105details
RTX 30708 GBFitsQ8_05.8 GB105details
RTX 3080 10GB10 GBFitsQ8_05.8 GB178details
RTX 3080 12GB12 GBFitsF169.3 GB113details
RTX 3090 Ti24 GBFitsF169.3 GB125details
RTX 4060 Ti 8GB8 GBFitsQ8_05.8 GB67details
RTX 4070 Super12 GBFitsF169.3 GB63details
RTX 4070 Ti12 GBFitsF169.3 GB63details
RTX 4070 Ti Super16 GBFitsF169.3 GB83details
RTX 408016 GBFitsF169.3 GB89details
RTX 4080 Super16 GBFitsF169.3 GB91details
RTX 50608 GBFitsQ8_05.8 GB105details
RTX 507012 GBFitsF169.3 GB83details
RTX 508016 GBFitsF169.3 GB119details
2x RTX 30902×24 GBFitsF169.3 GB116details
2x RTX 40902×24 GBFitsF169.3 GB125details
2x RTX 50902×32 GBFitsF169.3 GB223details
RTX A600048 GBFitsF169.3 GB95details
RTX 6000 Ada48 GBFitsF169.3 GB119details
RTX PRO 6000 Blackwell96 GBFitsF169.3 GB223details
L40S48 GBFitsF169.3 GB107details
A100 40GB40 GBFitsF169.3 GB193details
A100 80GB80 GBFitsF169.3 GB253details
H100 PCIe80 GBFitsF169.3 GB248details
H100 SXM80 GBFitsF169.3 GB416details
H200141 GBFitsF169.3 GB596details
RX 76008 GBFitsQ8_05.8 GB67details
RX 7700 XT12 GBFitsF169.3 GB54details
RX 7800 XT16 GBFitsF169.3 GB78details
RX 7900 GRE16 GBFitsF169.3 GB72details
RX 7900 XT20 GBFitsF169.3 GB99details
RX 907016 GBFitsF169.3 GB79details
RX 9070 XT16 GBFitsF169.3 GB79details
Ryzen AI Max+ 395 128GB128 GBFitsF169.3 GB32details
Arc A770 16GB16 GBFitsF169.3 GB70details
Mac M1 16GB16 GBFitsF169.3 GB8.4details
Mac M2 24GB24 GBFitsF169.3 GB12details
Mac M3 24GB24 GBFitsF169.3 GB12details
Mac M4 16GB16 GBFitsF169.3 GB15details
Mac M4 32GB32 GBFitsF169.3 GB15details
Mac M1 Pro 32GB32 GBFitsF169.3 GB25details
Mac M3 Pro 36GB36 GBFitsF169.3 GB19details
Mac M4 Pro 24GB24 GBFitsF169.3 GB34details
Mac M4 Pro 64GB64 GBFitsF169.3 GB34details
Mac M1 Max 64GB64 GBFitsF169.3 GB50details
Mac M2 Max 96GB96 GBFitsF169.3 GB50details
Mac M3 Max 128GB128 GBFitsF169.3 GB50details
Mac M4 Max 64GB64 GBFitsF169.3 GB68details
Mac M2 Ultra 192GB192 GBFitsF169.3 GB99details
Mac M3 Ultra 256GB256 GBFitsF169.3 GB102details

Questions

How much VRAM does Qwen3 4B Instruct 2507 need?

About 4.1 GB at 4-bit (Q4_K_M) with an 8K context: 2.3 GB of weights (real GGUF file size), 1.1 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 5.8 GB.

What is the smallest GPU that runs Qwen3 4B Instruct 2507?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 4060 (8 GB).

Can I run Qwen3 4B Instruct 2507 on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M1 16GB or larger, using the default macOS GPU memory limit.

Similar models

Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.

Qwen3 4B

4.0B · 4.1 GB at 4-bit

Qwen2.5 VL 3B Instruct

3.8B · 2.8 GB at 4-bit

Qwen3 VL 4B Instruct

4.4B · 4.1 GB at 4-bit

granite vision 4.1 4b

4.0B · 3.2 GB at 4-bit

Phi 3.5 vision instruct

4.1B · 5.8 GB at 4-bit

Phi 4 mini instruct

3.8B · 4.1 GB at 4-bit

Model facts

Hugging Face repoQwen/Qwen3-4B-Instruct-2507
Parameters4,022,468,096 (4.0B)
Layers / hidden size36 / 2560
AttentionGQA; 32 heads, 8 KV heads, head dim 128
KV cache per token (FP16)144.0 KiB
Max context262,144 tokens
Licenseapache-2.0
Released on HFAug 5, 2025
Downloads (30d) / likes2,729,704 / 1,057

Sources: Hugging Face model API · config.json (Qwen/Qwen3-4B-Instruct-2507) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.