RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can I run DeepSeek R1 Distill Qwen 32B locally?

DeepSeek R1 Distill Qwen 32B by deepseek-ai has 33B parameters. At 4-bit with an 8K context it needs about 21.2 GB of GPU memory. It fits on 48 of 65 GPUs and Macs we track. See all models that fit in 24 GB of VRAM.

21.2 GBQ4_K_M · 8K context
RTX 5090smallest consumer GPU that fits
M3 Pro 36GBsmallest Mac that fits
128Kmax 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/DeepSeek-R1-Distill-Qwen-32B-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K) •rough13.2 GB14.2 GB20.2 GB44.2 GB
3-bit (Q3_K_M) •usable16.5 GB17.5 GB23.5 GB47.5 GB
4-bit (Q4_K_M) •good, most popular20.2 GB21.2 GB27.2 GB51.2 GB
5-bit (Q5_K_M) •very good23.4 GB24.4 GB30.4 GB54.4 GB
6-bit (Q6_K) •near lossless26.7 GB27.7 GB33.7 GB57.7 GB
8-bit (Q8_0) •effectively lossless34.1 GB35.1 GB41.1 GB65.1 GB
16-bit (F16/BF16) •original weights62.7 GB63.7 GB69.7 GB93.7 GB

Which GPUs and Macs run DeepSeek R1 Distill Qwen 32B?

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 GBNo———details
RTX 4060 Ti 16GB16 GBTightQ2_K14.2 GB23details
RTX 309024 GBFitsQ3_K_M17.5 GB59details
RTX 409024 GBFitsQ3_K_M17.5 GB63details
RTX 509032 GBFitsQ5_K_M24.4 GB77details
RTX 5070 Ti16 GBTightQ2_K14.2 GB73details
RX 7900 XTX24 GBFitsQ3_K_M17.5 GB60details
Mac M4 Pro 48GB48 GBFitsQ6_K27.7 GB10details
Mac M4 Max 128GB128 GBFitsF1663.7 GB8.3details
Mac M3 Ultra 512GB512 GBFitsF1663.7 GB12details
DGX Spark128 GBFitsF1663.7 GB4.2details
RTX 407012 GBNo———details
RTX 5060 Ti 16GB16 GBTightQ2_K14.2 GB36details
RTX 40608 GBNo———details
Arc B58012 GBNo———details
RTX 3060 Ti8 GBNo———details
RTX 30708 GBNo———details
RTX 3080 10GB10 GBNo———details
RTX 3080 12GB12 GBNo———details
RTX 3090 Ti24 GBFitsQ3_K_M17.5 GB63details
RTX 4060 Ti 8GB8 GBNo———details
RTX 4070 Super12 GBNo———details
RTX 4070 Ti12 GBNo———details
RTX 4070 Ti Super16 GBTightQ2_K14.2 GB55details
RTX 408016 GBTightQ2_K14.2 GB58details
RTX 4080 Super16 GBTightQ2_K14.2 GB60details
RTX 50608 GBNo———details
RTX 507012 GBNo———details
RTX 508016 GBTightQ2_K14.2 GB78details
2x RTX 30902×24 GBFitsQ8_035.1 GB27details
2x RTX 40902×24 GBFitsQ8_035.1 GB29details
2x RTX 50902×32 GBFitsQ8_035.1 GB51details
RTX A600048 GBFitsQ8_035.1 GB22details
RTX 6000 Ada48 GBFitsQ8_035.1 GB28details
RTX PRO 6000 Blackwell96 GBFitsF1663.7 GB27details
L40S48 GBFitsQ8_035.1 GB25details
A100 40GB40 GBFitsQ6_K27.7 GB58details
A100 80GB80 GBFitsF1663.7 GB31details
H100 PCIe80 GBFitsF1663.7 GB31details
H100 SXM80 GBFitsF1663.7 GB51details
H200141 GBFitsF1663.7 GB73details
RX 76008 GBNo———details
RX 7700 XT12 GBNo———details
RX 7800 XT16 GBTightQ2_K14.2 GB51details
RX 7900 GRE16 GBTightQ2_K14.2 GB47details
RX 7900 XT20 GBFitsQ2_K14.2 GB65details
RX 907016 GBTightQ2_K14.2 GB52details
RX 9070 XT16 GBTightQ2_K14.2 GB52details
Ryzen AI Max+ 395 128GB128 GBFitsF1663.7 GB3.9details
Arc A770 16GB16 GBTightQ2_K14.2 GB45details
Mac M1 16GB16 GBNo———details
Mac M2 24GB24 GBFitsQ2_K14.2 GB8.1details
Mac M3 24GB24 GBFitsQ2_K14.2 GB8.1details
Mac M4 16GB16 GBNo———details
Mac M4 32GB32 GBFitsQ3_K_M17.5 GB7.5details
Mac M1 Pro 32GB32 GBFitsQ3_K_M17.5 GB13details
Mac M3 Pro 36GB36 GBFitsQ4_K_M21.2 GB7.6details
Mac M4 Pro 24GB24 GBFitsQ2_K14.2 GB22details
Mac M4 Pro 64GB64 GBFitsQ8_035.1 GB7.8details
Mac M1 Max 64GB64 GBFitsQ8_035.1 GB11details
Mac M2 Max 96GB96 GBFitsF1663.7 GB6.1details
Mac M3 Max 128GB128 GBFitsF1663.7 GB6.1details
Mac M4 Max 64GB64 GBFitsQ8_035.1 GB16details
Mac M2 Ultra 192GB192 GBFitsF1663.7 GB12details
Mac M3 Ultra 256GB256 GBFitsF1663.7 GB12details

Questions

How much VRAM does DeepSeek R1 Distill Qwen 32B need?

About 21.2 GB at 4-bit (Q4_K_M) with an 8K context: 18.5 GB of weights (real GGUF file size), 2.0 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 35.1 GB.

What is the smallest GPU that runs DeepSeek R1 Distill Qwen 32B?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 5090 (32 GB).

Can I run DeepSeek R1 Distill Qwen 32B on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M3 Pro 36GB 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.

deepseek moe 16b base

16B · 12.5 GB at 4-bit

DeepSeek Coder V2 Lite Instruct

16B · 10.5 GB at 4-bit

DeepSeek V2 Lite

16B · 9.8 GB at 4-bit

Qwen3 32B

33B · 21.1 GB at 4-bit

GLM Z1 32B 0414

33B · 19.5 GB at 4-bit

GLM 4 32B 0414

33B · 19.5 GB at 4-bit

Model facts

Hugging Face repodeepseek-ai/DeepSeek-R1-Distill-Qwen-32B
Parameters32,763,876,352 (33B)
Layers / hidden size64 / 5120
AttentionGQA; 40 heads, 8 KV heads, head dim 128
KV cache per token (FP16)256.0 KiB
Max context131,072 tokens
Licensemit
Released on HFJan 20, 2025
Downloads (30d) / likes367,533 / 1,676

Sources: Hugging Face model API · config.json (deepseek-ai/DeepSeek-R1-Distill-Qwen-32B) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.