gpt-oss-20b-Derestricted GGUF size and VRAM requirements
MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF is a large language model with 20.91 billion parameters, built on the gpt-oss architecture. It has been downloaded 95,545 times.
gpt-oss-20b-Derestricted-GGUF is a Mixture-of-Experts model with 32 experts, of which 4 are active on each token. Routing only 4 of 32 experts makes it noticeably faster than a dense model of the same size — but every expert still has to be held in memory, so the numbers below are set by the full parameter count, not the active one. It is built from 24 transformer layers, a hidden size of 2,880, 64 attention heads. It uses a hybrid attention design: only 12 of its 24 layers keep a full KV cache, while the rest use linear attention with a constant-size state. Long contexts therefore cost far less memory than on a conventional model of this size.
To run MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF locally at a 4,096-token context, its quantized versions need between 12.13 GB (Q2_K, lowest quality) and 39.88 GB (GGUF, highest quality) of memory, weights plus KV cache and a system margin included. Context length drives that memory directly: at 4,096 tokens the KV cache for gpt-oss-20b-Derestricted-GGUF is about 0.09 GB, rising to roughly 3.0 GB at its full 131,072-token context. Shorter prompts free up memory for a higher-quality quantization.
For most users the best balance is Q6_K, needing about 21.56 GB. That means MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF fits entirely in the VRAM of a 16 GB GPU or larger, running fully on the GPU.
Available GGUF quantizations for MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF include Q2_K, Q3_K_M, Q3_K_L, Q4_K_M, Q5_K_M, Q6_K, GGUF. The model supports a native context length of up to 131,072 tokens; a longer context grows the KV cache and the memory needed.
GGUF file size and memory by quantization
Compare real GGUF weight sizes, estimated KV cache and total memory for Q4, Q5, Q8 and every quantization published in this repository.
| Quant. | Bits | Quality | Weights | KV | Total | Speed~ | Verdict |
|---|---|---|---|---|---|---|---|
| Q2_K | 4.62 | Good | 11.24 GB | 0.09 GB | 12.13 GB | 4.4 t/s | Offload |
| Q3_K_M | 4.94 | Good | 12.03 GB | 0.09 GB | 12.92 GB | 4.2 t/s | Offload |
| Q3_K_L | 5.1 | Very good | 12.42 GB | 0.09 GB | 13.31 GB | 4.0 t/s | Offload |
| Q4_K_M | 6.05 | Very good | 14.72 GB | 0.09 GB | 15.61 GB | 3.4 t/s | Offload |
| Q5_K_M | 6.46 | Very good | 15.73 GB | 0.09 GB | 16.63 GB | 3.2 t/s | Offload |
| Q6_K | 8.49 | Excellent | 20.67 GB | 0.09 GB | 21.56 GB | 2.4 t/s | Offload |
| GGUF | 16.01 | Excellent | 38.99 GB | 0.09 GB | 39.88 GB | — | Insufficient |
KV cache computed from the model's exact architecture. Speed is a rough estimate bounded by memory bandwidth.
Frequently asked questions
What kind of model is MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF?
MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF is a language model with 20.91 billion parameters, based on the gpt-oss architecture. It is distributed as GGUF files for local inference.
Is MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF a Mixture-of-Experts (MoE) model?
Yes. MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF is a Mixture-of-Experts model with 32 experts, of which 4 are activated per token. That makes it faster than a dense model of the same size, but all 32 experts must be loaded into memory, so the VRAM/RAM it needs is driven by the total parameter count, not the active one.
How does MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF handle long context?
MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF uses a hybrid attention design: only 12 of its 24 layers keep a full KV cache, and the rest use linear attention with a constant-size state. Long contexts cost far less memory than on a conventional model of this size.
How much VRAM do you need to run MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF?
You need about 15.61 GB of VRAM to run MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF entirely on the GPU using the Q4_K_M quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.
Can I run MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF on an 8 GB GPU?
Partially. MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF only fits on an 8 GB GPU by offloading part of it to system RAM (with Q6_K), which runs but is slower.
Can I run MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF on a 16 GB GPU?
Yes. With 16 GB of VRAM you can run MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF fully on the GPU using Q4_K_M (about 15.61 GB).
Can I run MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF on a 24 GB GPU?
Yes. With 24 GB of VRAM you can run MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF fully on the GPU using Q6_K (about 21.56 GB).
What context length does MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF support?
MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF supports a native context length of up to 131,072 tokens. A longer context grows the KV cache, so it increases the memory needed to run the model.
What is the best quantization for MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF?
For MaziyarPanahi/gpt-oss-20b-Derestricted-GGUF, a strong default is Q4_K_M, which needs about 15.61 GB and keeps most of the quality while roughly halving the memory versus 8-bit. With VRAM to spare, Q5_K_M or Q6_K add a little more quality; if you are tight on memory, a smaller quantization still runs. Pick the highest quantization that fits your VRAM.