RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can I run WeVisOne VL 4B locally?

WeVisOne VL 4B by tencent has 4.8B parameters. At 4-bit with an 8K context it needs about 4.3 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track. See all models that fit in 8 GB of VRAM.

4.3 GBQ4_K_M · 8K context
RTX 4060smallest consumer GPU that fits
M1 16GBsmallest Mac that fits
256Kmax 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 mradermacher/WeVisOne-VL-4B-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K) •rough2.9 GB3.5 GB6.9 GB20.4 GB
3-bit (Q3_K_M) •usable3.3 GB3.9 GB7.3 GB20.8 GB
4-bit (Q4_K_M) •good, most popular3.8 GB4.3 GB7.7 GB21.2 GB
5-bit (Q5_K_M) •very good4.2 GB4.8 GB8.1 GB21.6 GB
6-bit (Q6_K) •near lossless4.6 GB5.2 GB8.6 GB22.1 GB
8-bit (Q8_0) •effectively lossless5.6 GB6.2 GB9.6 GB23.1 GB
16-bit (F16/BF16) •original weights9.5 GB10.0 GB13.4 GB26.9 GB

Which GPUs and Macs run WeVisOne VL 4B?

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 GBFitsF1610.0 GB41details
RTX 4060 Ti 16GB16 GBFitsF1610.0 GB33details
RTX 309024 GBFitsF1610.0 GB106details
RTX 409024 GBFitsF1610.0 GB114details
RTX 509032 GBFitsF1610.0 GB203details
RTX 5070 Ti16 GBFitsF1610.0 GB101details
RX 7900 XTX24 GBFitsF1610.0 GB109details
Mac M4 Pro 48GB48 GBFitsF1610.0 GB31details
Mac M4 Max 128GB128 GBFitsF1610.0 GB62details
Mac M3 Ultra 512GB512 GBFitsF1610.0 GB93details
DGX Spark128 GBFitsF1610.0 GB31details
RTX 407012 GBFitsF1610.0 GB57details
RTX 5060 Ti 16GB16 GBFitsF1610.0 GB51details
RTX 40608 GBFitsQ8_06.2 GB58details
Arc B58012 GBFitsF1610.0 GB52details
RTX 3060 Ti8 GBFitsQ8_06.2 GB95details
RTX 30708 GBFitsQ8_06.2 GB95details
RTX 3080 10GB10 GBFitsQ8_06.2 GB162details
RTX 3080 12GB12 GBFitsF1610.0 GB103details
RTX 3090 Ti24 GBFitsF1610.0 GB114details
RTX 4060 Ti 8GB8 GBFitsQ8_06.2 GB61details
RTX 4070 Super12 GBFitsF1610.0 GB57details
RTX 4070 Ti12 GBFitsF1610.0 GB57details
RTX 4070 Ti Super16 GBFitsF1610.0 GB76details
RTX 408016 GBFitsF1610.0 GB81details
RTX 4080 Super16 GBFitsF1610.0 GB83details
RTX 50608 GBFitsQ8_06.2 GB95details
RTX 507012 GBFitsF1610.0 GB76details
RTX 508016 GBFitsF1610.0 GB109details
2x RTX 30902×24 GBFitsF1610.0 GB106details
2x RTX 40902×24 GBFitsF1610.0 GB114details
2x RTX 50902×32 GBFitsF1610.0 GB203details
RTX A600048 GBFitsF1610.0 GB87details
RTX 6000 Ada48 GBFitsF1610.0 GB109details
RTX PRO 6000 Blackwell96 GBFitsF1610.0 GB203details
L40S48 GBFitsF1610.0 GB98details
A100 40GB40 GBFitsF1610.0 GB176details
A100 80GB80 GBFitsF1610.0 GB231details
H100 PCIe80 GBFitsF1610.0 GB226details
H100 SXM80 GBFitsF1610.0 GB379details
H200141 GBFitsF1610.0 GB543details
RX 76008 GBFitsQ8_06.2 GB61details
RX 7700 XT12 GBFitsF1610.0 GB49details
RX 7800 XT16 GBFitsF1610.0 GB71details
RX 7900 GRE16 GBFitsF1610.0 GB65details
RX 7900 XT20 GBFitsF1610.0 GB91details
RX 907016 GBFitsF1610.0 GB72details
RX 9070 XT16 GBFitsF1610.0 GB72details
Ryzen AI Max+ 395 128GB128 GBFitsF1610.0 GB29details
Arc A770 16GB16 GBFitsF1610.0 GB63details
Mac M1 16GB16 GBFitsQ8_06.2 GB14details
Mac M2 24GB24 GBFitsF1610.0 GB11details
Mac M3 24GB24 GBFitsF1610.0 GB11details
Mac M4 16GB16 GBFitsQ8_06.2 GB26details
Mac M4 32GB32 GBFitsF1610.0 GB14details
Mac M1 Pro 32GB32 GBFitsF1610.0 GB23details
Mac M3 Pro 36GB36 GBFitsF1610.0 GB17details
Mac M4 Pro 24GB24 GBFitsF1610.0 GB31details
Mac M4 Pro 64GB64 GBFitsF1610.0 GB31details
Mac M1 Max 64GB64 GBFitsF1610.0 GB45details
Mac M2 Max 96GB96 GBFitsF1610.0 GB45details
Mac M3 Max 128GB128 GBFitsF1610.0 GB45details
Mac M4 Max 64GB64 GBFitsF1610.0 GB62details
Mac M2 Ultra 192GB192 GBFitsF1610.0 GB91details
Mac M3 Ultra 256GB256 GBFitsF1610.0 GB93details

Questions

How much VRAM does WeVisOne VL 4B need?

About 4.3 GB at 4-bit (Q4_K_M) with an 8K context: 2.5 GB of weights (real GGUF file size), 1.1 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 6.2 GB.

What is the smallest GPU that runs WeVisOne VL 4B?

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 WeVisOne VL 4B 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.

Similar models

Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.

WeVisOne VL 2B

2.4B · 2.8 GB at 4-bit

Qwen3.8 27B DSpine b16

2.0B · 2.2 GB at 4-bit

HunyuanOCR

1.1B · 2.0 GB at 4-bit

Molmo2 4B

4.9B · 4.3 GB at 4-bit

Mage VL Cowork

4.7B · 4.3 GB at 4-bit

Qianfan OCR

4.7B · 4.4 GB at 4-bit

Model facts

Hugging Face repotencent/WeVisOne-VL-4B
Parameters4,830,581,248 (4.8B)
Layers / hidden size36 / 2560
AttentionGQA; 32 heads, 8 KV heads, head dim 128
KV cache per token (FP16)144.0 KiB
Max context262,144 tokens
Licenseapache-2.0
Released on HFOct 22, 2026
Downloads (30d) / likes26,116 / 127

Sources: Hugging Face model API · config.json (tencent/WeVisOne-VL-4B) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.