ME

Mellum2 12B A2.5B Thinking Abliterated

MellumGGUF by josephmayoUpdated 2026-06-02
Download on Hugging Face
Parameters12.1B
Context128K
Layers28
Architecturemellum
Smallest quant8.2 GB
Quick answer

Mellum2 12B A2.5B Thinking Abliterated needs about 11.8 GB of memory at Q6_K (a 10.9 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 3 quants, from 8.2 GB (Q4_K_L) to 12.9 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 1.9 GB.

Mellum2 12B A2.5B Thinking Abliterated – a reasoning‑focused MoE model

Mellum2 12B A2.5B Thinking Abliterated is a 12.1B parameter model with a 2.5B active MoE. It is trained for reasoning and includes refusal abliterations across all experts. The model supports a 131072 token context window.

Good at

  • Reasoning with chain‑of‑thought
  • Large context handling
  • MoE‑based refusal ablation

Limits

  • No vision support
  • License not specified
  • Requires large VRAM for full precision

Best for

  • Complex reasoning tasks
  • Long‑form content generation
  • AI safety research

Who it’s for. Researchers and developers needing a reasoning‑oriented MoE model with built‑in refusal ablation. Suitable for experimentation and safety studies.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K_L8.2 GB 9.2 GB
Too big Get file →
Q6_K10.9 GB 11.8 GB
Too big Get file →
Q8_012.9 GB 13.9 GB
Too big Get file →

For 8 GB VRAM, use Q4_K_L (8.24 GB file, 9.2 GB memory at 8k context). For 12 GB VRAM, Q6_K (10.88 GB file, 11.8 GB memory). For 16 GB VRAM, Q8_0 (12.93 GB file, 13.9 GB memory). Memory estimates are approximate.

What GPU do you need for Mellum2 12B A2.5B Thinking 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: 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: the best quant that fits is Q6_K (10.9 GB file, about 11.8 GB in use).
  • 16 GB VRAM: the best quant that fits is Q8_0 (12.9 GB file, about 13.9 GB in use).

Which Mellum2 12B A2.5B Thinking Abliterated quant should you choose?

Q6_K (10.9 GB) is the middle option in this repository. Very close to the original quality for most uses, at a clearly larger size than 4-bit.

If memory is tight, the smallest file is Q4_K_L at 8.2 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturemellum
Parameters12.1B (approximate, mixture of experts)
Active parameters2.5B
Layers28
Trained context131,072 tokens
KV cache (fp16)56 KB per token

Source

Quantized byjosephmayo
FamilyMellum
Downloads (HF)603
Last updated2026-06-02

Frequently asked questions about Mellum2 12B A2.5B Thinking Abliterated

How much VRAM does Mellum2 12B A2.5B Thinking Abliterated need?

About 11.8 GB at Q6_K with an 8K context window. That is the 10.9 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 Mellum2 12B A2.5B Thinking Abliterated on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 9.2 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Mellum2 12B A2.5B Thinking Abliterated GGUF?

It depends on the quant: Q4_K_L is 8.2 GB, Q6_K is 10.9 GB, Q8_0 is 12.9 GB.

What context length does Mellum2 12B A2.5B Thinking 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 0.5 GB at 8K and 1.9 GB at 32K.

Where can I download Mellum2 12B A2.5B Thinking Abliterated GGUF?

The files are on Hugging Face in the josephmayo/Mellum2-12B-A2.5B-Thinking-Abliterated-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Can I run Mellum2 12B A2.5B on 8 GB VRAM?

Yes, using the Q4_K_L quantization which requires about 9.2 GB of VRAM at 8k context.

Does this model support vision inputs?

No, the model has no vision capabilities.

What is the context length of this model?

It supports up to 131072 tokens.

Other GGUFs of Mellum2-12B-A2.5B-Thinking-Abliterated

Similar models

Browse all →

More from josephmayo

All by josephmayo →

More Mellum models

All Mellum →

Other medium models

All medium models →