LI

Ling 3.0 flash VL MTP ROCmFP4

LingGGUF by kingjones777Updated 2026-09-22
Download on Hugging Face
Parameters124B
Context128K
Layers42
Architecturebailingmoe3
Smallest quant66.5 GB
Quick answer

Ling 3.0 flash VL MTP ROCmFP4 needs about 83.9 GB of memory at Q4_0 (a 67.9 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 4 quants, from 66.5 GB (Q4_0) to 72.1 GB (Q4_0). Longer context needs more: at 32K tokens the KV cache alone adds about 62.0 GB.

Ling 3.0 flash VL MTP ROCmFP4 – 124B multimodal model for vision and text

Ling 3.0 flash VL MTP ROCmFP4 is a 124‑billion‑parameter multimodal model built on Ling‑3.0‑flash‑VL. It supports image and video inputs, a 256K‑token context window, and uses a sparse MoE architecture with 5.5B active parameters per token. The model is quantized to ROCmFP4 for AMD GPUs.

Good at

  • Vision and text reasoning
  • Long‑context multimodal tasks
  • Efficient inference with sparse MoE

Limits

  • No text‑only quantization variants
  • Requires AMD ROCm support
  • No explicit language list provided

Best for

  • Multimodal question answering
  • Video analysis and editing
  • Agentic workflows with vision

Who it’s for. Developers building vision‑enabled agents on AMD hardware. The model is ideal for applications needing long‑context reasoning and native image/video understanding.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_066.5 GB 82.5 GB
Too big Get file →
Q4_067.9 GB 83.9 GB
Too big Get file →
Q4_068.1 GB 84.1 GB
Too big Get file →
Q4_072.1 GB 88.1 GB
Too big Get file →

For 8 GB VRAM, use the Q4_0 file of 66.49 GB (est. 82.5 GB memory at 8 k context). For 12 GB VRAM, the 67.93 GB file (83.9 GB memory) is suitable. Larger VRAM allows the 68.14 GB or 72.12 GB files with higher memory estimates.

What GPU do you need for Ling 3.0 flash VL MTP ROCmFP4?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: no quant fits at an 8K context.
  • 32 GB VRAM: no quant fits at an 8K context.
  • 48 GB VRAM: no quant fits at an 8K context.

Which Ling 3.0 flash VL MTP ROCmFP4 quant should you choose?

Q4_0 (67.9 GB) is the middle option in this repository. The usual sweet spot between file size and quality, and the quant most people start with.

If memory is tight, the smallest file is Q4_0 at 66.5 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturebailingmoe3
Parameters124B (approximate, mixture of experts)
Layers42
Trained context131,072 tokens
KV cache (fp16)1848 KB per token

Source

Quantized bykingjones777
FamilyLing
Licensemit
Downloads (HF)516
Last updated2026-09-22

Frequently asked questions about Ling 3.0 flash VL MTP ROCmFP4

How much VRAM does Ling 3.0 flash VL MTP ROCmFP4 need?

About 83.9 GB at Q4_0 with an 8K context window. That is the 67.9 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Ling 3.0 flash VL MTP ROCmFP4 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 82.5 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Ling 3.0 flash VL MTP ROCmFP4 GGUF?

It depends on the quant: Q4_0 is 66.5 GB, Q4_0 is 67.9 GB, Q4_0 is 68.1 GB, Q4_0 is 72.1 GB.

What context length does Ling 3.0 flash VL MTP ROCmFP4 support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 15.5 GB at 8K and 62.0 GB at 32K.

Where can I download Ling 3.0 flash VL MTP ROCmFP4 GGUF?

The files are on Hugging Face in the kingjones777/Ling-3.0-flash-VL-MTP-ROCmFP4-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Ling 3.0 flash VL MTP ROCmFP4 use?

Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Ling 3.0 flash VL MTP ROCmFP4 on 8 GB VRAM?

Use the 66.49 GB Q4_0 file; it requires approximately 82.5 GB memory at 8 k context, so 8 GB VRAM is insufficient for full model loading.

Does this model support text-only inference?

The provided quantizations are for multimodal use; no text‑only variant is listed.

What hardware is required to run this model?

AMD GPUs with ROCm support, such as the Strix Halo (gfx1151), are required for ROCmFP4 quantization.

Other GGUFs of Ling-3.0-flash-VL

Similar models

Browse all →

More from kingjones777

All by kingjones777 →

More Ling models

All Ling →

Other large models

All large models →