RunsOnMy ModelsGPUs & MacsAPI pricesVRAM calculatorNew

Can I run Ling 3.0 tiny locally?

Ling 3.0 tiny by inclusionAI has 7.9B parameters (1.4B active per token, 128 experts, 8 used per token). At 4-bit with an 8K context it needs about 5.5 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track.

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

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from bartowski/Ling-3.0-tiny-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K) •rough3.6 GB3.7 GB4.3 GB6.9 GB
3-bit (Q3_K_M) •usable4.3 GB4.4 GB5.1 GB7.6 GB
4-bit (Q4_K_M) •good, most popular5.4 GB5.5 GB6.1 GB8.7 GB
5-bit (Q5_K_M) •very good6.1 GB6.2 GB6.9 GB9.4 GB
6-bit (Q6_K) •near lossless7.2 GB7.3 GB7.9 GB10.4 GB
8-bit (Q8_0) •effectively lossless8.6 GB8.7 GB9.4 GB11.9 GB
16-bit (F16/BF16) •original weights15.5 GB15.6 GB16.3 GB18.8 GB

Which GPUs and Macs run Ling 3.0 tiny?

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 GBFitsQ8_08.7 GB245details
RTX 4060 Ti 16GB16 GBFitsQ8_08.7 GB196details
RTX 309024 GBFitsF1615.6 GB338details
RTX 409024 GBFitsF1615.6 GB364details
RTX 509032 GBFitsF1615.6 GB648details
RTX 5070 Ti16 GBFitsQ8_08.7 GB609details
RX 7900 XTX24 GBFitsF1615.6 GB347details
Mac M4 Pro 48GB48 GBFitsF1615.6 GB99details
Mac M4 Max 128GB128 GBFitsF1615.6 GB197details
Mac M3 Ultra 512GB512 GBFitsF1615.6 GB296details
DGX Spark128 GBFitsF1615.6 GB99details
RTX 407012 GBFitsQ8_08.7 GB342details
RTX 5060 Ti 16GB16 GBFitsQ8_08.7 GB304details
RTX 40608 GBFitsQ5_K_M6.2 GB272details
Arc B58012 GBFitsQ8_08.7 GB310details
RTX 3060 Ti8 GBFitsQ5_K_M6.2 GB448details
RTX 30708 GBFitsQ5_K_M6.2 GB448details
RTX 3080 10GB10 GBFitsQ6_K7.3 GB635details
RTX 3080 12GB12 GBFitsQ8_08.7 GB620details
RTX 3090 Ti24 GBFitsF1615.6 GB364details
RTX 4060 Ti 8GB8 GBFitsQ5_K_M6.2 GB288details
RTX 4070 Super12 GBFitsQ8_08.7 GB342details
RTX 4070 Ti12 GBFitsQ8_08.7 GB342details
RTX 4070 Ti Super16 GBFitsQ8_08.7 GB456details
RTX 408016 GBFitsQ8_08.7 GB487details
RTX 4080 Super16 GBFitsQ8_08.7 GB500details
RTX 50608 GBFitsQ5_K_M6.2 GB448details
RTX 507012 GBFitsQ8_08.7 GB456details
RTX 508016 GBFitsQ8_08.7 GB652details
2x RTX 30902×24 GBFitsF1615.6 GB338details
2x RTX 40902×24 GBFitsF1615.6 GB364details
2x RTX 50902×32 GBFitsF1615.6 GB648details
RTX A600048 GBFitsF1615.6 GB278details
RTX 6000 Ada48 GBFitsF1615.6 GB347details
RTX PRO 6000 Blackwell96 GBFitsF1615.6 GB648details
L40S48 GBFitsF1615.6 GB312details
A100 40GB40 GBFitsF1615.6 GB562details
A100 80GB80 GBFitsF1615.6 GB737details
H100 PCIe80 GBFitsF1615.6 GB723details
H100 SXM80 GBFitsF1615.6 GB1211details
H200141 GBFitsF1615.6 GB1735details
RX 76008 GBFitsQ5_K_M6.2 GB288details
RX 7700 XT12 GBFitsQ8_08.7 GB293details
RX 7800 XT16 GBFitsQ8_08.7 GB424details
RX 7900 GRE16 GBFitsQ8_08.7 GB391details
RX 7900 XT20 GBFitsF1615.6 GB289details
RX 907016 GBFitsQ8_08.7 GB435details
RX 9070 XT16 GBFitsQ8_08.7 GB435details
Ryzen AI Max+ 395 128GB128 GBFitsF1615.6 GB93details
Arc A770 16GB16 GBFitsQ8_08.7 GB380details
Mac M1 16GB16 GBFitsQ8_08.7 GB46details
Mac M2 24GB24 GBFitsQ8_08.7 GB68details
Mac M3 24GB24 GBFitsQ8_08.7 GB68details
Mac M4 16GB16 GBFitsQ8_08.7 GB82details
Mac M4 32GB32 GBFitsF1615.6 GB43details
Mac M1 Pro 32GB32 GBFitsF1615.6 GB72details
Mac M3 Pro 36GB36 GBFitsF1615.6 GB54details
Mac M4 Pro 24GB24 GBFitsQ8_08.7 GB185details
Mac M4 Pro 64GB64 GBFitsF1615.6 GB99details
Mac M1 Max 64GB64 GBFitsF1615.6 GB145details
Mac M2 Max 96GB96 GBFitsF1615.6 GB145details
Mac M3 Max 128GB128 GBFitsF1615.6 GB145details
Mac M4 Max 64GB64 GBFitsF1615.6 GB197details
Mac M2 Ultra 192GB192 GBFitsF1615.6 GB289details
Mac M3 Ultra 256GB256 GBFitsF1615.6 GB296details

Prefer an API?

inclusionAI: Ling 3.0 Tiny is on OpenRouter at $0.015 per 1M input tokens and $0.045 per 1M output tokens. Price details and history →

Questions

How much VRAM does Ling 3.0 tiny need?

About 5.5 GB at 4-bit (Q4_K_M) with an 8K context: 4.6 GB of weights (real GGUF file size), 0.2 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 8.7 GB.

What is the smallest GPU that runs Ling 3.0 tiny?

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 Ling 3.0 tiny 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 repoinclusionAI/Ling-3.0-tiny
Parameters7,893,392,800 (7.9B) · ~1.4B active (computed from config)
Layers / hidden size24 / 1536
AttentionMLA; 16 heads, 16 KV heads, head dim 128
KV cache per token (FP16)27.0 KiB
Max context131,072 tokens
Licensemit
Released on HFAug 10, 2026
Downloads (30d) / likes13,680 / 588

Sources: Hugging Face model API · config.json (inclusionAI/Ling-3.0-tiny) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.