RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

Can I run deepseek coder 7b instruct v1.5 locally?

deepseek coder 7b instruct v1.5 by deepseek-ai has 6.9B parameters. At 4-bit with an 8K context it needs about 8.3 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track.

8.3 GBQ4_K_M · 8K context
RTX 3080 10GBsmallest consumer GPU that fits
M1 16GBsmallest Mac that fits
4Kmax 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 mradermacher/deepseek-coder-7b-instruct-v1.5-i1-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx
2-bit (Q2_K) •rough5.0 GB
3-bit (Q3_K_M) •usable5.7 GB
4-bit (Q4_K_M) •good, most popular6.4 GB
5-bit (Q5_K_M) •very good7.1 GB
6-bit (Q6_K) •near lossless7.8 GB
8-bit (Q8_0)effectively lossless9.3 GB
16-bit (F16/BF16)original weights15.3 GB

Which GPUs and Macs run deepseek coder 7b instruct v1.5?

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 GBFitsQ6_K9.6 GB63details
RTX 4060 Ti 16GB16 GBFitsQ8_011.2 GB39details
RTX 309024 GBFitsF1617.2 GB68details
RTX 409024 GBFitsF1617.2 GB73details
RTX 509032 GBFitsF1617.2 GB130details
RTX 5070 Ti16 GBFitsQ8_011.2 GB122details
RX 7900 XTX24 GBFitsF1617.2 GB69details
Mac M4 Pro 48GB48 GBFitsF1617.2 GB20details
Mac M4 Max 128GB128 GBFitsF1617.2 GB40details
Mac M3 Ultra 512GB512 GBFitsF1617.2 GB59details
DGX Spark128 GBFitsF1617.2 GB20details
RTX 407012 GBFitsQ6_K9.6 GB89details
RTX 5060 Ti 16GB16 GBFitsQ8_011.2 GB61details
RTX 40608 GBTightQ3_K_M7.6 GB79details
Arc B58012 GBFitsQ6_K9.6 GB80details
RTX 3060 Ti8 GBTightQ3_K_M7.6 GB129details
RTX 30708 GBTightQ3_K_M7.6 GB129details
RTX 3080 10GB10 GBFitsQ4_K_M8.3 GB180details
RTX 3080 12GB12 GBFitsQ6_K9.6 GB161details
RTX 3090 Ti24 GBFitsF1617.2 GB73details
RTX 4060 Ti 8GB8 GBTightQ3_K_M7.6 GB83details
RTX 4070 Super12 GBFitsQ6_K9.6 GB89details
RTX 4070 Ti12 GBFitsQ6_K9.6 GB89details
RTX 4070 Ti Super16 GBFitsQ8_011.2 GB92details
RTX 408016 GBFitsQ8_011.2 GB98details
RTX 4080 Super16 GBFitsQ8_011.2 GB100details
RTX 50608 GBTightQ3_K_M7.6 GB129details
RTX 507012 GBFitsQ6_K9.6 GB118details
RTX 508016 GBFitsQ8_011.2 GB131details
2x RTX 30902×24 GBFitsF1617.2 GB68details
2x RTX 40902×24 GBFitsF1617.2 GB73details
2x RTX 50902×32 GBFitsF1617.2 GB130details
RTX A600048 GBFitsF1617.2 GB56details
RTX 6000 Ada48 GBFitsF1617.2 GB69details
RTX PRO 6000 Blackwell96 GBFitsF1617.2 GB130details
L40S48 GBFitsF1617.2 GB63details
A100 40GB40 GBFitsF1617.2 GB113details
A100 80GB80 GBFitsF1617.2 GB148details
H100 PCIe80 GBFitsF1617.2 GB145details
H100 SXM80 GBFitsF1617.2 GB242details
H200141 GBFitsF1617.2 GB347details
RX 76008 GBTightQ3_K_M7.6 GB83details
RX 7700 XT12 GBFitsQ6_K9.6 GB76details
RX 7800 XT16 GBFitsQ8_011.2 GB85details
RX 7900 GRE16 GBFitsQ8_011.2 GB78details
RX 7900 XT20 GBFitsQ8_011.2 GB109details
RX 907016 GBFitsQ8_011.2 GB87details
RX 9070 XT16 GBFitsQ8_011.2 GB87details
Ryzen AI Max+ 395 128GB128 GBFitsF1617.2 GB19details
Arc A770 16GB16 GBFitsQ8_011.2 GB76details
Mac M1 16GB16 GBFitsQ6_K9.6 GB12details
Mac M2 24GB24 GBFitsQ8_011.2 GB14details
Mac M3 24GB24 GBFitsQ8_011.2 GB14details
Mac M4 16GB16 GBFitsQ6_K9.6 GB21details
Mac M4 32GB32 GBFitsF1617.2 GB8.7details
Mac M1 Pro 32GB32 GBFitsF1617.2 GB14details
Mac M3 Pro 36GB36 GBFitsF1617.2 GB11details
Mac M4 Pro 24GB24 GBFitsQ8_011.2 GB37details
Mac M4 Pro 64GB64 GBFitsF1617.2 GB20details
Mac M1 Max 64GB64 GBFitsF1617.2 GB29details
Mac M2 Max 96GB96 GBFitsF1617.2 GB29details
Mac M3 Max 128GB128 GBFitsF1617.2 GB29details
Mac M4 Max 64GB64 GBFitsF1617.2 GB40details
Mac M2 Ultra 192GB192 GBFitsF1617.2 GB58details
Mac M3 Ultra 256GB256 GBFitsF1617.2 GB59details

Questions

How much VRAM does deepseek coder 7b instruct v1.5 need?

About 8.3 GB at 4-bit (Q4_K_M) with an 8K context: 3.9 GB of weights (real GGUF file size), 3.8 GB of KV cache and 0.6 GB of runtime overhead. At 8-bit it needs about 11.2 GB.

What is the smallest GPU that runs deepseek coder 7b instruct v1.5?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 3080 10GB (10 GB).

Can I run deepseek coder 7b instruct v1.5 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-7b-instruct-v1.5
Parameters6,910,365,696 (6.9B)
Layers / hidden size30 / 4096
AttentionMHA; 32 heads, 32 KV heads, head dim 128
KV cache per token (FP16)480.0 KiB
Max context4,096 tokens
Licenseother
Released on HFJan 25, 2024
Downloads (30d) / likes92,732 / 195

Sources: Hugging Face model API · config.json (deepseek-ai/deepseek-coder-7b-instruct-v1.5) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.