SmallThinker 21B A3B Instruct MXFP4 MOE needs about 13.8 GB of memory at MXFP4_MOE (a 12.5 GB file) with an 8K context window, so it fits on a 16 GB GPU.
SmallThinker 21B A3B Instruct MXFP4 MOE is a 21.5B-parameter Smallthinker model in GGUF format, quantized by noctrex. Its trained context length is 16K tokens. The repository has 1 quant file, from 12.5 GB (MXFP4_MOE) to 12.5 GB (MXFP4_MOE).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| MXFP4_MOE | 12.5 GB | 13.8 GB | Too big | Get file → |
What GPU do you need for SmallThinker 21B A3B Instruct MXFP4 MOE?
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: the best quant that fits is MXFP4_MOE (12.5 GB file, about 13.8 GB in use).
Which SmallThinker 21B A3B Instruct MXFP4 MOE quant should you choose?
MXFP4_MOE (12.5 GB) is the middle option in this repository. Lower bit counts mean smaller files and more quality loss.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about SmallThinker 21B A3B Instruct MXFP4 MOE
How much VRAM does SmallThinker 21B A3B Instruct MXFP4 MOE need?
About 13.8 GB at MXFP4_MOE with an 8K context window. That is the 12.5 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 SmallThinker 21B A3B Instruct MXFP4 MOE on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 13.8 GB, so you would need a larger GPU or CPU offloading.
What is the file size of SmallThinker 21B A3B Instruct MXFP4 MOE GGUF?
It depends on the quant: MXFP4_MOE is 12.5 GB.
What context length does SmallThinker 21B A3B Instruct MXFP4 MOE support?
The GGUF header lists a trained context length of 16,384 tokens (16K). The KV cache grows with the context you use: about 0.9 GB at 8K and 1.7 GB at 16K.
Where can I download SmallThinker 21B A3B Instruct MXFP4 MOE GGUF?
The files are on Hugging Face in the noctrex/SmallThinker-21B-A3B-Instruct-MXFP4_MOE-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.