GP

gpt oss 20b DFlash

gpt-ossGGUF by EntityDeletrUpdated 2026-07-01
Download on Hugging Face
Parameters785M
Context128K
Layers8
Architecturedflash
Smallest quant0.6 GB
Quick answer

gpt oss 20b DFlash needs about 1.2 GB of memory at its main file (a 0.6 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 2 quants, from 0.6 GB to 1.6 GB. Longer context needs more: at 32K tokens the KV cache alone adds about 0.5 GB.

gpt‑oss‑20b‑DFlash – lightweight drafter for speculative decoding

gpt‑oss‑20b‑DFlash is a 0.8B‑parameter drafter built on the DFlash architecture. It is designed for efficient, high‑quality parallel drafting and must be paired with the target model openai/gpt‑oss‑20b. The model is released under the Apache‑2.0 license.

Good at

  • Efficient speculative drafting
  • Block diffusion for parallel drafting
  • Lightweight drafter component

Limits

  • No vision support
  • Requires target model openai/gpt‑oss‑20b
  • No mixture of experts

Best for

  • Parallel drafting in inference pipelines
  • Speed‑up speculative decoding
  • Low‑memory drafting tasks

Who it’s for. Developers building inference pipelines that need a fast drafter component. The model is suitable for users who already have the target gpt‑oss‑20b model and want to accelerate generation via speculative decoding.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.6 GB 1.2 GB
Fits Get file →
—1.6 GB 2.2 GB
Fits Get file →

For 8 GB VRAM, use the 0.6 GB file; it requires approximately 1.2 GB of VRAM at 8 k context. For larger memory budgets, the 1.58 GB file uses about 2.2 GB of VRAM. Both files are quantized to an unknown quant type and are compatible with llama.cpp.

What GPU do you need for gpt oss 20b DFlash?

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: the best quant that fits is the main file (1.6 GB file, about 2.2 GB in use).

Specifications

Architecture from the GGUF header

Architecturedflash
Parameters785M
Layers8
Trained context131,072 tokens
KV cache (fp16)16 KB per token

Source

Quantized byEntityDeletr
Familygpt-oss
Licenseapache-2.0
Downloads (HF)793
Last updated2026-07-01

Frequently asked questions about gpt oss 20b DFlash

How much VRAM does gpt oss 20b DFlash need?

About 1.2 GB at its main file with an 8K context window. That is the 0.6 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 gpt oss 20b DFlash on an 8 GB GPU?

Yes, with the smaller files: the 1.6 GB file needs about 2.2 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of gpt oss 20b DFlash GGUF?

It depends on the quant: file is 0.6 GB, file is 1.6 GB.

What context length does gpt oss 20b DFlash support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 0.1 GB at 8K and 0.5 GB at 32K.

Where can I download gpt oss 20b DFlash GGUF?

The files are on Hugging Face in the EntityDeletr/gpt-oss-20b-DFlash-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 gpt oss 20b DFlash 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 this model on 8 GB VRAM?

Yes, the 0.6 GB file requires about 1.2 GB of VRAM at 8 k context, which fits within an 8 GB GPU.

Does this model support vision?

No, the model has no vision capabilities.

What license is the model released under?

The model is released under the Apache‑2.0 license.

Similar models

Browse all →

More from EntityDeletr

All by EntityDeletr →

More gpt-oss models

All gpt-oss →

Other tiny models

All tiny models →