RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

Can I run deepseek coder 1.3b instruct locally?

deepseek coder 1.3b instruct by deepseek-ai has 1.3B parameters. At 4-bit with an 8K context it needs about 2.8 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track.

2.8 GBQ4_K_M · 8K context
RTX 4060smallest consumer GPU that fits
M1 16GBsmallest Mac that fits
16Kmax context
MHAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from TheBloke/deepseek-coder-1.3b-instruct-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx
2-bit (Q2_K) •rough1.8 GB2.5 GB
3-bit (Q3_K_M) •usable1.9 GB2.6 GB
4-bit (Q4_K_M) •good, most popular2.0 GB2.8 GB
5-bit (Q5_K_M) •very good2.1 GB2.9 GB
6-bit (Q6_K) •near lossless2.3 GB3.1 GB
8-bit (Q8_0) •effectively lossless2.5 GB3.3 GB
16-bit (F16/BF16)original weights3.7 GB4.5 GB

Which GPUs and Macs run deepseek coder 1.3b instruct?

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 GBFitsF164.5 GB134details
RTX 4060 Ti 16GB16 GBFitsF164.5 GB107details
RTX 309024 GBFitsF164.5 GB348details
RTX 409024 GBFitsF164.5 GB374details
RTX 509032 GBFitsF164.5 GB665details
RTX 5070 Ti16 GBFitsF164.5 GB333details
RX 7900 XTX24 GBFitsF164.5 GB356details
Mac M4 Pro 48GB48 GBFitsF164.5 GB101details
Mac M4 Max 128GB128 GBFitsF164.5 GB203details
Mac M3 Ultra 512GB512 GBFitsF164.5 GB304details
DGX Spark128 GBFitsF164.5 GB101details
RTX 407012 GBFitsF164.5 GB187details
RTX 5060 Ti 16GB16 GBFitsF164.5 GB166details
RTX 40608 GBFitsF164.5 GB101details
Arc B58012 GBFitsF164.5 GB169details
RTX 3060 Ti8 GBFitsF164.5 GB166details
RTX 30708 GBFitsF164.5 GB166details
RTX 3080 10GB10 GBFitsF164.5 GB282details
RTX 3080 12GB12 GBFitsF164.5 GB339details
RTX 3090 Ti24 GBFitsF164.5 GB374details
RTX 4060 Ti 8GB8 GBFitsF164.5 GB107details
RTX 4070 Super12 GBFitsF164.5 GB187details
RTX 4070 Ti12 GBFitsF164.5 GB187details
RTX 4070 Ti Super16 GBFitsF164.5 GB250details
RTX 408016 GBFitsF164.5 GB266details
RTX 4080 Super16 GBFitsF164.5 GB273details
RTX 50608 GBFitsF164.5 GB166details
RTX 507012 GBFitsF164.5 GB250details
RTX 508016 GBFitsF164.5 GB356details
2x RTX 30902×24 GBFitsF164.5 GB348details
2x RTX 40902×24 GBFitsF164.5 GB374details
2x RTX 50902×32 GBFitsF164.5 GB665details
RTX A600048 GBFitsF164.5 GB285details
RTX 6000 Ada48 GBFitsF164.5 GB356details
RTX PRO 6000 Blackwell96 GBFitsF164.5 GB665details
L40S48 GBFitsF164.5 GB321details
A100 40GB40 GBFitsF164.5 GB577details
A100 80GB80 GBFitsF164.5 GB757details
H100 PCIe80 GBFitsF164.5 GB743details
H100 SXM80 GBFitsF164.5 GB1244details
H200141 GBFitsF164.5 GB1782details
RX 76008 GBFitsF164.5 GB107details
RX 7700 XT12 GBFitsF164.5 GB160details
RX 7800 XT16 GBFitsF164.5 GB232details
RX 7900 GRE16 GBFitsF164.5 GB214details
RX 7900 XT20 GBFitsF164.5 GB297details
RX 907016 GBFitsF164.5 GB238details
RX 9070 XT16 GBFitsF164.5 GB238details
Ryzen AI Max+ 395 128GB128 GBFitsF164.5 GB95details
Arc A770 16GB16 GBFitsF164.5 GB208details
Mac M1 16GB16 GBFitsF164.5 GB25details
Mac M2 24GB24 GBFitsF164.5 GB37details
Mac M3 24GB24 GBFitsF164.5 GB37details
Mac M4 16GB16 GBFitsF164.5 GB45details
Mac M4 32GB32 GBFitsF164.5 GB45details
Mac M1 Pro 32GB32 GBFitsF164.5 GB74details
Mac M3 Pro 36GB36 GBFitsF164.5 GB56details
Mac M4 Pro 24GB24 GBFitsF164.5 GB101details
Mac M4 Pro 64GB64 GBFitsF164.5 GB101details
Mac M1 Max 64GB64 GBFitsF164.5 GB149details
Mac M2 Max 96GB96 GBFitsF164.5 GB149details
Mac M3 Max 128GB128 GBFitsF164.5 GB149details
Mac M4 Max 64GB64 GBFitsF164.5 GB203details
Mac M2 Ultra 192GB192 GBFitsF164.5 GB297details
Mac M3 Ultra 256GB256 GBFitsF164.5 GB304details

Questions

How much VRAM does deepseek coder 1.3b instruct need?

About 2.8 GB at 4-bit (Q4_K_M) with an 8K context: 0.8 GB of weights (real GGUF file size), 1.5 GB of KV cache and 0.5 GB of runtime overhead. At 8-bit it needs about 3.3 GB.

What is the smallest GPU that runs deepseek coder 1.3b instruct?

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 deepseek coder 1.3b instruct 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.

Model facts

Hugging Face repodeepseek-ai/deepseek-coder-1.3b-instruct
Parameters1,346,471,936 (1.3B)
Layers / hidden size24 / 2048
AttentionMHA; 16 heads, 16 KV heads, head dim 128
KV cache per token (FP16)192.0 KiB
Max context16,384 tokens
Licenseother
Released on HFOct 29, 2023
Downloads (30d) / likes34,550 / 214

Sources: Hugging Face model API · config.json (deepseek-ai/deepseek-coder-1.3b-instruct) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.