Can I run Ling 3.0 flash Sante locally?
Ling 3.0 flash Sante by inclusionAI has 127B parameters (5.6B active per token, 512 experts, 8 used per token). At 4-bit with an 8K context it needs about 75.9 GB of GPU memory. It fits on 17 of 65 GPUs and Macs we track.
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-flash-Sante-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.
| Quant | Quality | 4K ctx | 8K ctx | 32K ctx | 128K ctx |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | rough | 47.0 GB | 47.2 GB | 48.3 GB | 52.7 GB |
| 3-bit (Q3_K_M) • | usable | 58.2 GB | 58.3 GB | 59.5 GB | 63.9 GB |
| 4-bit (Q4_K_M) • | good, most popular | 75.8 GB | 75.9 GB | 77.0 GB | 81.5 GB |
| 5-bit (Q5_K_M) • | very good | 91.8 GB | 92.0 GB | 93.1 GB | 97.5 GB |
| 6-bit (Q6_K) • | near lossless | 105 GB | 105 GB | 106 GB | 110 GB |
| 8-bit (Q8_0) • | effectively lossless | 127 GB | 127 GB | 128 GB | 133 GB |
| 16-bit (F16/BF16) • | original weights | 238 GB | 239 GB | 240 GB | 244 GB |
Which GPUs and Macs run Ling 3.0 flash Sante?
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).
Questions
How much VRAM does Ling 3.0 flash Sante need?
About 75.9 GB at 4-bit (Q4_K_M) with an 8K context: 74.9 GB of weights (real GGUF file size), 0.4 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 127 GB.
What is the smallest GPU that runs Ling 3.0 flash Sante?
No single consumer GPU in our table fits it at Q4_K_M with 8K context; the smallest option that does is the RTX PRO 6000 Blackwell (96 GB).
Can I run Ling 3.0 flash Sante on a Mac?
Yes. At Q4_K_M with 8K context it fits on a Mac M3 Max 128GB 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.
Ling 3.0 flash
127B · 73.5 GB at 4-bitLLaDA2.1 mini
16B · 10.3 GB at 4-bitLing 3.0 tiny
7.9B · 5.5 GB at 4-bitOrnith 1.5 122B
125B · 72.3 GB at 4-bitgpt oss 120b
117B · 59.5 GB at 4-bitgpt oss safeguard 120b
117B · 59.5 GB at 4-bitModel facts
| Hugging Face repo | inclusionAI/Ling-3.0-flash-Sante |
| Parameters | 127,486,405,600 (127B) · ~5.6B active (computed from config) |
| Layers / hidden size | 42 / 2560 |
| Attention | MLA; 32 heads, 32 KV heads, head dim 128 |
| KV cache per token (FP16) | 47.3 KiB |
| Max context | 262,144 tokens |
| License | mit |
| Released on HF | Oct 19, 2026 |
| Downloads (30d) / likes | 2,818 / 86 |
Sources: Hugging Face model API · config.json (inclusionAI/Ling-3.0-flash-Sante) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.