Gpt Oss 120b Abliterated needs about 89.0 GB of memory at Q4_K_M (a 87.9 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 25 quants, from 0.2 GB to 124.2 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 2.4 GB.
120B GPT‑OSS model with MoE and abliterated safety experts
Gpt Oss 120b Abliterated is a 120 billion‑parameter GPT‑OSS variant that uses mixture‑of‑experts and abliterated safety experts. It is released under Apache‑2.0 and supports a 131 072‑token context window.
Good at
- Large‑scale language modeling
- MoE‑based inference
- High‑token‑limit contexts
Limits
- Requires large VRAM for full‑precision MoE
- No vision or multimodal support
- License is Apache‑2.0
Best for
- Research on large‑scale MoE models
- Benchmarking refusal‑suppression techniques
- Running long‑form text generation
Who it’s for. Researchers and developers needing a 120 billion‑parameter GPT‑OSS model with MoE and safety‑suppressed experts. The model is suitable for environments that can handle the memory demands of the full MoE architecture.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.2 GB | 1.3 GB | Fits | Get file → | |
| MXFP4_MOE | 63.4 GB | 64.5 GB | Too big | Get file → | |
| IQ1_M | 66.1 GB | 67.2 GB | Too big | Get file → | |
| IQ1_S | 66.1 GB | 67.2 GB | Too big | Get file → | |
| IQ2_XXS | 66.1 GB | 67.2 GB | Too big | Get file → | |
| IQ2_XS | 66.1 GB | 67.2 GB | Too big | Get file → | |
| Q3_K_S | 66.1 GB | 67.3 GB | Too big | Get file → | |
| IQ2_M | 66.2 GB | 67.3 GB | Too big | Get file → | |
| IQ2_S | 66.2 GB | 67.3 GB | Too big | Get file → | |
| IQ3_S | 66.2 GB | 67.3 GB | Too big | Get file → | |
| IQ3_XS | 66.2 GB | 67.3 GB | Too big | Get file → | |
| IQ3_XXS | 66.2 GB | 67.3 GB | Too big | Get file → | |
| Q2_K | 66.2 GB | 67.3 GB | Too big | Get file → | |
| IQ4_XS | 66.2 GB | 67.3 GB | Too big | Get file → | |
| Q4_0 | 66.5 GB | 67.6 GB | Too big | Get file → | |
| Q2_K_S | 66.6 GB | 67.7 GB | Too big | Get file → | |
| IQ3_M | 66.7 GB | 67.8 GB | Too big | Get file → | |
| Q3_K_M | 71.1 GB | 72.2 GB | Too big | Get file → | |
| Q3_K_L | 73.4 GB | 74.5 GB | Too big | Get file → | |
| Q4_1 | 73.5 GB | 74.6 GB | Too big | Get file → | |
| Q4_K_S | 80.9 GB | 82.0 GB | Too big | Get file → | |
| Q4_K_M | 87.9 GB | 89.0 GB | Too big | Get file → | |
| Q5_K_S | 88.0 GB | 89.1 GB | Too big | Get file → | |
| Q5_K_M | 93.9 GB | 95.0 GB | Too big | Get file → | |
| Q6_K | 124.2 GB | 125.3 GB | Too big | Get file → |
For 8 GB VRAM, use the 0.16 GB unknown quant, which needs about 1.3 GB memory at 8 k context. For 16 GB VRAM, the Q4_0 quant (66.47 GB file) requires roughly 67.6 GB memory at 8 k context, so it is not feasible. The only practical quant for 8–16 GB VRAM is the 0.16 GB file. For larger GPUs, choose Q4_K_M (87.85 GB file) or Q5_K_M (93.94 GB file).
What GPU do you need for Gpt Oss 120b Abliterated?
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 (0.2 GB file, about 1.3 GB in use).
- 6 GB VRAM: the best quant that fits is the main file (0.2 GB file, about 1.3 GB in use).
- 8 GB VRAM: the best quant that fits is the main file (0.2 GB file, about 1.3 GB in use).
- 12 GB VRAM: the best quant that fits is the main file (0.2 GB file, about 1.3 GB in use).
- 16 GB VRAM: the best quant that fits is the main file (0.2 GB file, about 1.3 GB in use).
- 24 GB VRAM: the best quant that fits is the main file (0.2 GB file, about 1.3 GB in use).
- 32 GB VRAM: the best quant that fits is the main file (0.2 GB file, about 1.3 GB in use).
- 48 GB VRAM: the best quant that fits is the main file (0.2 GB file, about 1.3 GB in use).
Which Gpt Oss 120b Abliterated quant should you choose?
Q4_K_M is the usual starting point for Gpt Oss 120b Abliterated: it balances file size and quality well. Here it is a 87.9 GB file.
If memory is tight, the smallest file is the first one in the table at 0.2 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Gpt Oss 120b Abliterated
How much VRAM does Gpt Oss 120b Abliterated need?
About 89.0 GB at Q4_K_M with an 8K context window. That is the 87.9 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 Gpt Oss 120b Abliterated on an 8 GB GPU?
Yes, with the smaller files: the 0.2 GB file needs about 1.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Gpt Oss 120b Abliterated GGUF?
It depends on the quant: file is 0.2 GB, MXFP4_MOE is 63.4 GB, IQ1_M is 66.1 GB, IQ1_S is 66.1 GB, IQ2_XXS is 66.1 GB, IQ2_XS is 66.1 GB and 19 more.
What context length does Gpt Oss 120b Abliterated support?
The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 0.6 GB at 8K and 2.4 GB at 32K.
Where can I download Gpt Oss 120b Abliterated GGUF?
The files are on Hugging Face in the mradermacher/gpt-oss-120b-abliterated-i1-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 120b Abliterated 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?
Only the 0.16 GB unknown quant can fit, using about 1.3 GB memory at 8 k context. All other quants require far more VRAM.
Does this model support vision or multimodal inputs?
No, the model has no vision or multimodal capabilities.
What license applies to this model?
The model is released under the Apache‑2.0 license.