RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

Can I run HunyuanOCR locally?

HunyuanOCR by tencent has 1.1B parameters. At 4-bit with an 8K context it needs about 2.0 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track.

2.0 GBQ4_K_M · 8K context
RTX 4060smallest consumer GPU that fits
M1 16GBsmallest Mac that fits
128Kmax context
GQAattention type

GGUF Q8_0 file was 49% of the size expected from the parameter count, so this page uses size estimates instead of real file sizes.

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K)rough1.4 GB1.8 GB4.0 GB13.0 GB
3-bit (Q3_K_M)usable1.5 GB1.9 GB4.1 GB13.1 GB
4-bit (Q4_K_M)good, most popular1.6 GB2.0 GB4.3 GB13.3 GB
5-bit (Q5_K_M)very good1.7 GB2.1 GB4.4 GB13.4 GB
6-bit (Q6_K)near lossless1.9 GB2.2 GB4.5 GB13.5 GB
8-bit (Q8_0)effectively lossless2.1 GB2.5 GB4.7 GB13.7 GB
16-bit (F16/BF16)original weights3.1 GB3.5 GB5.7 GB14.7 GB

Which GPUs and Macs run HunyuanOCR?

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 GBFitsF163.5 GB161details
RTX 4060 Ti 16GB16 GBFitsF163.5 GB129details
RTX 309024 GBFitsF163.5 GB418details
RTX 409024 GBFitsF163.5 GB450details
RTX 509032 GBFitsF163.5 GB800details
RTX 5070 Ti16 GBFitsF163.5 GB400details
RX 7900 XTX24 GBFitsF163.5 GB429details
Mac M4 Pro 48GB48 GBFitsF163.5 GB122details
Mac M4 Max 128GB128 GBFitsF163.5 GB244details
Mac M3 Ultra 512GB512 GBFitsF163.5 GB366details
DGX Spark128 GBFitsF163.5 GB122details
RTX 407012 GBFitsF163.5 GB225details
RTX 5060 Ti 16GB16 GBFitsF163.5 GB200details
RTX 40608 GBFitsF163.5 GB121details
Arc B58012 GBFitsF163.5 GB204details
RTX 3060 Ti8 GBFitsF163.5 GB200details
RTX 30708 GBFitsF163.5 GB200details
RTX 3080 10GB10 GBFitsF163.5 GB339details
RTX 3080 12GB12 GBFitsF163.5 GB407details
RTX 3090 Ti24 GBFitsF163.5 GB450details
RTX 4060 Ti 8GB8 GBFitsF163.5 GB129details
RTX 4070 Super12 GBFitsF163.5 GB225details
RTX 4070 Ti12 GBFitsF163.5 GB225details
RTX 4070 Ti Super16 GBFitsF163.5 GB300details
RTX 408016 GBFitsF163.5 GB320details
RTX 4080 Super16 GBFitsF163.5 GB329details
RTX 50608 GBFitsF163.5 GB200details
RTX 507012 GBFitsF163.5 GB300details
RTX 508016 GBFitsF163.5 GB429details
2x RTX 30902×24 GBFitsF163.5 GB418details
2x RTX 40902×24 GBFitsF163.5 GB450details
2x RTX 50902×32 GBFitsF163.5 GB800details
RTX A600048 GBFitsF163.5 GB343details
RTX 6000 Ada48 GBFitsF163.5 GB429details
RTX PRO 6000 Blackwell96 GBFitsF163.5 GB800details
L40S48 GBFitsF163.5 GB386details
A100 40GB40 GBFitsF163.5 GB694details
A100 80GB80 GBFitsF163.5 GB910details
H100 PCIe80 GBFitsF163.5 GB893details
H100 SXM80 GBFitsF163.5 GB1496details
H200141 GBFitsF163.5 GB2143details
RX 76008 GBFitsF163.5 GB129details
RX 7700 XT12 GBFitsF163.5 GB193details
RX 7800 XT16 GBFitsF163.5 GB279details
RX 7900 GRE16 GBFitsF163.5 GB257details
RX 7900 XT20 GBFitsF163.5 GB357details
RX 907016 GBFitsF163.5 GB286details
RX 9070 XT16 GBFitsF163.5 GB286details
Ryzen AI Max+ 395 128GB128 GBFitsF163.5 GB114details
Arc A770 16GB16 GBFitsF163.5 GB250details
Mac M1 16GB16 GBFitsF163.5 GB30details
Mac M2 24GB24 GBFitsF163.5 GB45details
Mac M3 24GB24 GBFitsF163.5 GB45details
Mac M4 16GB16 GBFitsF163.5 GB54details
Mac M4 32GB32 GBFitsF163.5 GB54details
Mac M1 Pro 32GB32 GBFitsF163.5 GB89details
Mac M3 Pro 36GB36 GBFitsF163.5 GB67details
Mac M4 Pro 24GB24 GBFitsF163.5 GB122details
Mac M4 Pro 64GB64 GBFitsF163.5 GB122details
Mac M1 Max 64GB64 GBFitsF163.5 GB179details
Mac M2 Max 96GB96 GBFitsF163.5 GB179details
Mac M3 Max 128GB128 GBFitsF163.5 GB179details
Mac M4 Max 64GB64 GBFitsF163.5 GB244details
Mac M2 Ultra 192GB192 GBFitsF163.5 GB357details
Mac M3 Ultra 256GB256 GBFitsF163.5 GB366details

Questions

How much VRAM does HunyuanOCR need?

About 2.0 GB at 4-bit (Q4_K_M) with an 8K context: 0.6 GB of weights (estimated), 0.8 GB of KV cache and 0.6 GB of runtime overhead. At 8-bit it needs about 2.5 GB.

What is the smallest GPU that runs HunyuanOCR?

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 HunyuanOCR 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 repotencent/HunyuanOCR
Parameters1,119,925,744 (1.1B)
Layers / hidden size24 / 1024
AttentionGQA; 16 heads, 8 KV heads, head dim 128
KV cache per token (FP16)96.0 KiB
Max context131,072 tokens
Licenseother
Released on HFNov 18, 2025
Downloads (30d) / likes640,777 / 847

Sources: Hugging Face model API · config.json (tencent/HunyuanOCR) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.