RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

Can I run Olmo 3 7B Instruct locally?

Olmo 3 7B Instruct by allenai has 7.3B parameters. At 4-bit with an 8K context it needs about 7.3 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track.

7.3 GBQ4_K_M · 8K context
RTX 3080 10GBsmallest consumer GPU that fits
M1 16GBsmallest Mac that fits
64Kmax context
MHA + sliding windowattention 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/Olmo-3-7B-Instruct-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx
2-bit (Q2_K) •rough5.3 GB5.8 GB8.8 GB
3-bit (Q3_K_M) •usable6.0 GB6.5 GB9.5 GB
4-bit (Q4_K_M) •good, most popular6.8 GB7.3 GB10.3 GB
5-bit (Q5_K_M) •very good7.4 GB7.9 GB10.9 GB
6-bit (Q6_K) •near lossless8.2 GB8.7 GB11.7 GB
8-bit (Q8_0) •effectively lossless9.8 GB10.3 GB13.3 GB
16-bit (F16/BF16) •original weights16.2 GB16.7 GB19.7 GB

Which GPUs and Macs run Olmo 3 7B 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 GBFitsQ6_K8.7 GB60details
RTX 4060 Ti 16GB16 GBFitsQ8_010.3 GB37details
RTX 309024 GBFitsF1616.7 GB64details
RTX 409024 GBFitsF1616.7 GB69details
RTX 509032 GBFitsF1616.7 GB123details
RTX 5070 Ti16 GBFitsQ8_010.3 GB115details
RX 7900 XTX24 GBFitsF1616.7 GB66details
Mac M4 Pro 48GB48 GBFitsF1616.7 GB19details
Mac M4 Max 128GB128 GBFitsF1616.7 GB37details
Mac M3 Ultra 512GB512 GBFitsF1616.7 GB56details
DGX Spark128 GBFitsF1616.7 GB19details
RTX 407012 GBFitsQ6_K8.7 GB84details
RTX 5060 Ti 16GB16 GBFitsQ8_010.3 GB58details
RTX 40608 GBFitsQ3_K_M6.5 GB74details
Arc B58012 GBFitsQ6_K8.7 GB76details
RTX 3060 Ti8 GBFitsQ3_K_M6.5 GB123details
RTX 30708 GBFitsQ3_K_M6.5 GB123details
RTX 3080 10GB10 GBFitsQ5_K_M7.9 GB146details
RTX 3080 12GB12 GBFitsQ6_K8.7 GB152details
RTX 3090 Ti24 GBFitsF1616.7 GB69details
RTX 4060 Ti 8GB8 GBFitsQ3_K_M6.5 GB79details
RTX 4070 Super12 GBFitsQ6_K8.7 GB84details
RTX 4070 Ti12 GBFitsQ6_K8.7 GB84details
RTX 4070 Ti Super16 GBFitsQ8_010.3 GB87details
RTX 408016 GBFitsQ8_010.3 GB92details
RTX 4080 Super16 GBFitsQ8_010.3 GB95details
RTX 50608 GBFitsQ3_K_M6.5 GB123details
RTX 507012 GBFitsQ6_K8.7 GB112details
RTX 508016 GBFitsQ8_010.3 GB124details
2x RTX 30902×24 GBFitsF1616.7 GB64details
2x RTX 40902×24 GBFitsF1616.7 GB69details
2x RTX 50902×32 GBFitsF1616.7 GB123details
RTX A600048 GBFitsF1616.7 GB53details
RTX 6000 Ada48 GBFitsF1616.7 GB66details
RTX PRO 6000 Blackwell96 GBFitsF1616.7 GB123details
L40S48 GBFitsF1616.7 GB59details
A100 40GB40 GBFitsF1616.7 GB107details
A100 80GB80 GBFitsF1616.7 GB140details
H100 PCIe80 GBFitsF1616.7 GB137details
H100 SXM80 GBFitsF1616.7 GB229details
H200141 GBFitsF1616.7 GB329details
RX 76008 GBFitsQ3_K_M6.5 GB79details
RX 7700 XT12 GBFitsQ6_K8.7 GB72details
RX 7800 XT16 GBFitsQ8_010.3 GB80details
RX 7900 GRE16 GBFitsQ8_010.3 GB74details
RX 7900 XT20 GBFitsF1616.7 GB55details
RX 907016 GBFitsQ8_010.3 GB82details
RX 9070 XT16 GBFitsQ8_010.3 GB82details
Ryzen AI Max+ 395 128GB128 GBFitsF1616.7 GB18details
Arc A770 16GB16 GBFitsQ8_010.3 GB72details
Mac M1 16GB16 GBFitsQ6_K8.7 GB11details
Mac M2 24GB24 GBFitsQ8_010.3 GB13details
Mac M3 24GB24 GBFitsQ8_010.3 GB13details
Mac M4 16GB16 GBFitsQ6_K8.7 GB20details
Mac M4 32GB32 GBFitsF1616.7 GB8.2details
Mac M1 Pro 32GB32 GBFitsF1616.7 GB14details
Mac M3 Pro 36GB36 GBFitsF1616.7 GB10details
Mac M4 Pro 24GB24 GBFitsQ8_010.3 GB35details
Mac M4 Pro 64GB64 GBFitsF1616.7 GB19details
Mac M1 Max 64GB64 GBFitsF1616.7 GB27details
Mac M2 Max 96GB96 GBFitsF1616.7 GB27details
Mac M3 Max 128GB128 GBFitsF1616.7 GB27details
Mac M4 Max 64GB64 GBFitsF1616.7 GB37details
Mac M2 Ultra 192GB192 GBFitsF1616.7 GB55details
Mac M3 Ultra 256GB256 GBFitsF1616.7 GB56details

Questions

How much VRAM does Olmo 3 7B Instruct need?

About 7.3 GB at 4-bit (Q4_K_M) with an 8K context: 4.2 GB of weights (real GGUF file size), 2.5 GB of KV cache and 0.6 GB of runtime overhead. At 8-bit it needs about 10.3 GB.

What is the smallest GPU that runs Olmo 3 7B Instruct?

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 Olmo 3 7B 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 repoallenai/Olmo-3-7B-Instruct
Parameters7,298,011,136 (7.3B)
Layers / hidden size32 / 4096
AttentionMHA + sliding window; 32 heads, 32 KV heads, head dim 128; sliding window 4096
KV cache per token (FP16)512.0 KiB (sliding layers capped at the window)
Max context65,536 tokens
Licenseapache-2.0
Released on HFNov 19, 2025
Downloads (30d) / likes131,479 / 160

Sources: Hugging Face model API · config.json (allenai/Olmo-3-7B-Instruct) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.