Ling 3.1 Flash, GGUF
Ling 3.1 Flash is a native hybrid reasoning Mixture-of-Experts model designed for general-purpose agents, office productivity, and software engineering.
This repository holds Vultr's GGUF quantizations of it. They are hybrid quantizations: rather than applying one precision to the whole network, each layer's precision is chosen from a measured importance score, so the bits go where they change the output most.
Model Overview
| Attribute | Value |
|---|---|
| Base model | inclusionAI/Ling-3.1-flash |
| Architecture | bailingmoe3, Mixture-of-Experts |
| Parameters | 560B total |
| Active parameters | 25B |
| Layers | 72 |
| Hidden size | 4,096 |
| Attention heads | 64 |
| Key-value heads | 64 |
| Experts | 8 of 512 selected per token plus a shared expert |
| Context length | 1M-token |
| Vocabulary | 157,184 |
| Attention layout | Kimi Delta Attention and gated MLA layers alternate in a 7:1 ratio |
| License | MIT |
| Attribute | Value |
|---|---|
| Base model | inclusionAI/Ling-3.1-flash |
| Architecture | bailingmoe3, Mixture-of-Experts |
| Parameters | 560B total |
| Active parameters | 25B |
| Layers | 72 |
| Hidden size | 4,096 |
| Attention heads | 64 |
| Key-value heads | 64 |
| Experts | 8 of 512 selected per token plus a shared expert |
| Context length | 1M-token |
| Vocabulary | 157,184 |
| Attention layout | Kimi Delta Attention and gated MLA layers alternate in a 7:1 ratio |
| License | MIT |
Quantizations
| File | Type | Size | KLD vs bf16 |
|---|---|---|---|
| Ling-3.1-flash-Q5_K_M-*.gguf (3 parts) | Q5_K_M | 439.69 GB | 0.0127 |
| Ling-3.1-flash-Q4_K_M-*.gguf (3 parts) | Q4_K_M | 374.12 GB | 0.0228 |
| Ling-3.1-flash-IQ3_M-*.gguf (2 parts) | IQ3_M | 270.02 GB | 0.0816 |
| Ling-3.1-flash-IQ3_S-*.gguf (2 parts) | IQ3_S | 267.37 GB | 0.0932 |
| File | Type | Size | KLD vs bf16 |
|---|---|---|---|
| Ling-3.1-flash-Q5_K_M-*.gguf (3 parts) | Q5_K_M | 439.69 GB | 0.0127 |
| Ling-3.1-flash-Q4_K_M-*.gguf (3 parts) | Q4_K_M | 374.12 GB | 0.0228 |
| Ling-3.1-flash-IQ3_M-*.gguf (2 parts) | IQ3_M | 270.02 GB | 0.0816 |
| Ling-3.1-flash-IQ3_S-*.gguf (2 parts) | IQ3_S | 267.37 GB | 0.0932 |
KL divergence measures how far a quantized model's next-token probabilities drift from those of the original bf16 model, averaged over every token of the evaluation text. The lower the value, the closer the quantized model is to the original; zero would mean identical output.
Evaluation
Every variant in this repository was evaluated against the original bf16 model for perplexity, KL divergence, top-token agreement, HellaSwag and WinoGrande, under identical conditions. The full results are maintained on the model's page in the Vultr Model Library:
https://docs.vultr.com/models/inclusionai-ling-31/ling-3.1-flash?tab=quantized
That page is the reference for these files. The summary above is an excerpt from it.
Run
llama-server -hf vultr/Ling-3.1-flash-GGUF:Q4_K_M
The model's chat template is embedded in every file, so no extra prompt formatting is needed.
The repository also includes mtp-Ling-3.1-flash-Q8_0.gguf, the model's multi-token prediction (MTP) head, for optional speculative decoding with any quant above:
llama-server -hf vultr/Ling-3.1-flash-GGUF:Q4_K_M --spec-type draft-mtp
License
These files are derived from inclusionAI/Ling-3.1-flash and are distributed under the same license, MIT. The model is the work of inclusionAI; Vultr produced the quantizations.