Ling 3.0 flash Sante
Ling 3.0 flash Sante is an open-weights language model from inclusionAI, released October 19, 2026. It has 127B parameters (mixture-of-experts, 8.6B active). The context window is 256K tokens.
Check it against your own PC →
Can you run Ling 3.0 flash Sante locally?
The smallest common GPU it runs on is 12 GB of VRAM (with 32 GB of RAM): IQ2_XXS, GPU + experts in RAM, roughly 37 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Too large | – | – | – | – |
| 8 GB VRAM | Too large | – | – | – | – |
| 12 GB VRAM | Runs · GPU + CPU experts | IQ2_XXS (heavy loss) | 11.2 GB | 23.0 GB | ~37 |
| 16 GB VRAM | Runs · GPU + CPU experts | IQ2_XXS (heavy loss) | 15.2 GB | 19.0 GB | ~44 |
| 24 GB VRAM | Runs · GPU + CPU experts | IQ2_M (heavy loss) | 23.2 GB | 21.2 GB | ~45 |
| 32 GB VRAM | Runs · GPU + CPU experts | IQ3_XXS (noticeable loss) | 31.2 GB | 19.0 GB | ~52 |
Ready-made quantizations on Hugging Face: AtomicChat/Ling-3.0-flash-Sante-GGUF · AtomicChat/Ling-3.0-flash-Sante-Uncensored-GGUF · bartowski/Ling-3.0-flash-Sante-GGUF · vultr/Ling-3.0-flash-Sante-GGUF · dicksondickson/Ling-3.0-flash-Sante-oQ4e-mtp-bf16-MLX · aitamedia/Ling-3.0-flash-Sante-FP4
Benchmark results
Epoch AI has not published results for Ling 3.0 flash Sante yet. New models are usually evaluated within days to weeks.
Links
Hugging Face: inclusionAI/Ling-3.0-flash-Sante
More from inclusionAI
Ling 3.1 flash 1019 · LBS 27B · Ling 3.0 tiny · Ling-1T · Ring-1T