inclusionAI LLaDA2.2 flash needs about 60.1 GB of memory at Q4_K_S (a 59.1 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 2 quants, from 59.1 GB (Q4_K_S) to 205.9 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 2.1 GB.
inclusionAI LLaDA2.2 flash – a 103B‑parameter diffusion LLM
inclusionAI LLaDA2.2 flash is a 103B‑parameter diffusion language model with a 128K context window. It uses Levenshtein Editing and block routing for efficient long‑context agentic workloads. The model is released under Apache‑2.0.
Good at
- Long‑context agentic tasks
- Levenshtein editing
- Block routing for efficient diffusion
Limits
- Requires diffuse‑cpp, not llama.cpp
- Large VRAM needed
- No vision support
Best for
- Agentic applications with long context
- Tool use with block routing
- Error correction via Levenshtein editing
Who it’s for. Developers building agentic systems that need a 128K context window and block‑diffusion inference. The model is suitable for research and production environments that can host large diffusion models.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K_S | 59.1 GB (2 parts) | 60.1 GB | Too big | Get file → | |
| F16 | 205.9 GB (5 parts) | 206.9 GB | Too big | Get file → |
The Q4_K_S quant is 59.09 GB and requires about 60.1 GB of VRAM at 8 k context, suitable for systems with around 60 GB VRAM. The F16 quant is 205.85 GB and needs about 206.9 GB of VRAM, suitable for high‑end servers.
What GPU do you need for inclusionAI LLaDA2.2 flash?
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 inclusionAI LLaDA2.2 flash quant should you choose?
Q4_K_S (59.1 GB) is the middle option in this repository. 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
Source
Frequently asked questions about inclusionAI LLaDA2.2 flash
How much VRAM does inclusionAI LLaDA2.2 flash need?
About 60.1 GB at Q4_K_S with an 8K context window. That is the 59.1 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 inclusionAI LLaDA2.2 flash on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 60.1 GB, so you would need a larger GPU or CPU offloading.
What is the file size of inclusionAI LLaDA2.2 flash GGUF?
It depends on the quant: Q4_K_S is 59.1 GB, F16 is 205.9 GB.
What context length does inclusionAI LLaDA2.2 flash support?
The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 0.5 GB at 8K and 2.1 GB at 32K.
Where can I download inclusionAI LLaDA2.2 flash GGUF?
The files are on Hugging Face in the Akicou/inclusionAI_LLaDA2.2-flash-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 inclusionAI LLaDA2.2 flash use?
Hugging Face lists the license as apache-2.0. 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 inclusionAI LLaDA2.2 flash on 8 GB VRAM?
No, the model requires more than 60 GB of VRAM for the smallest quant.
Does this model work with llama.cpp?
No, LLaDA2 is a block diffusion model and requires diffuse‑cpp.
What is the maximum context length?
The model supports a 128K context window.