clef flash heretic needs about 19.5 GB of memory at F16 (a 17.9 GB file) with an 8K context window, so it fits on a 24 GB GPU. Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.
Clef Flash Heretic – 9B multimodal decision model with suppressed refusals
Clef Flash Heretic is a 9B Qwen3.5‑based model that reads text, JSON, images, or video and outputs probabilities for each question option. It produces no free‑form text and is designed for decision‑making pipelines.
Good at
- Decision‑making from multimodal input
- Fast probability scoring for many options
- Suppresses refusal behavior
Limits
- No free‑text generation
- Requires 16 GB GPU for F16
- No other quantizations available yet
Best for
- Moderation triage
- Routing decisions
- Risk scoring
Who it’s for. Engineers building classification and decision pipelines who need reliable probability outputs without free‑text generation. Suitable for moderation triage, routing, and content classification tasks.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| F16 | 17.9 GB | 19.5 GB | Too big | Get file → |
The only published quant is F16 at 17.92 GB. It requires about 19.5 GB of VRAM at 8 k context, so a 16 GB GPU can run it with a small context window. For larger context, a higher‑capacity GPU is needed.
What GPU do you need for clef flash heretic?
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: the best quant that fits is F16 (17.9 GB file, about 19.5 GB in use).
Which clef flash heretic quant should you choose?
F16 (17.9 GB) is the middle option in this repository. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about clef flash heretic
How much VRAM does clef flash heretic need?
About 19.5 GB at F16 with an 8K context window. That is the 17.9 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 clef flash heretic on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 19.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of clef flash heretic GGUF?
It depends on the quant: F16 is 17.9 GB.
What context length does clef flash heretic support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.
Where can I download clef flash heretic GGUF?
The files are on Hugging Face in the saidutta69/clef-flash-heretic repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does clef flash heretic 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 Clef Flash Heretic on a 16 GB GPU?
Yes, the F16 quant uses about 19.5 GB of VRAM at 8 k context, which fits a 16 GB GPU with a small context window.
Does this model generate text?
No, it outputs only probabilities for each option and does not produce free‑form text.
What vision inputs are supported?
It accepts images via PIL and video frames as arrays, but requires pillow for image handling.