Ornith 1.5 35B A3B BigBang MTP needs about 22.3 GB of memory at Q4_K_M (a 21.2 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 6 quants, from 2.0 GB (Q8_0) to 36.9 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 2.7 GB.
Ornith 1.5 35B A3B BigBang MTP – a merged Qwen3.6‑based MoE model with a trained MTP head
Ornith 1.5 35B A3B BigBang MTP merges Ornith‑1.5‑35B‑A3B and BigBang‑v1 over Qwen3.6‑35B‑A3B. It replaces the untrained MTP placeholder with a trained head, enabling speculative decoding. The model is a mixture of experts with 1.86 b parameters and 3 b active parameters.
Good at
- Large context handling
- Mixture‑of‑experts inference
- Speculative decoding with MTP
Limits
- Requires recent llama.cpp (b10507+).
- No vision support in main GGUFs; draft must be added separately.
- License MIT with attribution to parents.
Best for
- Local inference on GPUs with 8 k context
- Speed‑critical generation with draft
- Research on MoE architectures
Who it’s for. Developers and researchers who need a local Qwen3.6‑based model with a trained MTP head for faster generation. The model is suitable for those who run llama.cpp or compatible tools and want to experiment with speculative decoding.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q8_0 | 2.0 GB | 3.2 GB | Fits | Get file → | |
| Q4_K_M | 21.2 GB | 22.3 GB | Too big | Get file → | |
| Q4_K_M | 21.7 GB | 22.9 GB | Too big | Get file → | |
| Q5_K_M | 24.7 GB | 25.9 GB | Too big | Get file → | |
| Q6_K | 28.5 GB | 29.7 GB | Too big | Get file → | |
| Q8_0 | 36.9 GB | 38.1 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_M (file size 21.17 GB, memory 22.3 GB at 8 k context). For 16 GB VRAM, Q8_0 (file size 1.99 GB, memory 3.2 GB) provides the best speed. If you need a larger draft, download mtpdraft‑Q8_0.gguf (2 GB).
What GPU do you need for Ornith 1.5 35B A3B BigBang MTP?
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 Q8_0 (2.0 GB file, about 3.2 GB in use).
- 6 GB VRAM: the best quant that fits is Q8_0 (2.0 GB file, about 3.2 GB in use).
- 8 GB VRAM: the best quant that fits is Q8_0 (2.0 GB file, about 3.2 GB in use).
- 12 GB VRAM: the best quant that fits is Q8_0 (2.0 GB file, about 3.2 GB in use).
- 16 GB VRAM: the best quant that fits is Q8_0 (2.0 GB file, about 3.2 GB in use).
- 24 GB VRAM: the best quant that fits is Q4_K_M (21.7 GB file, about 22.9 GB in use).
- 32 GB VRAM: the best quant that fits is Q6_K (28.5 GB file, about 29.7 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (36.9 GB file, about 38.1 GB in use).
Which Ornith 1.5 35B A3B BigBang MTP quant should you choose?
Q4_K_M is the usual starting point for Ornith 1.5 35B A3B BigBang MTP: it balances file size and quality well. Here it is a 21.2 GB file.
The Q8_0 file is 2.0 GB, about 0.1 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is Q8_0 at 2.0 GB, with the quality trade-offs that come with fewer bits. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Ornith 1.5 35B A3B BigBang MTP
How much VRAM does Ornith 1.5 35B A3B BigBang MTP need?
About 22.3 GB at Q4_K_M with an 8K context window. That is the 21.2 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 Ornith 1.5 35B A3B BigBang MTP on an 8 GB GPU?
Yes, with Q8_0: the 2.0 GB file needs about 3.2 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Ornith 1.5 35B A3B BigBang MTP GGUF?
It depends on the quant: Q8_0 is 2.0 GB, Q4_K_M is 21.2 GB, Q4_K_M is 21.7 GB, Q5_K_M is 24.7 GB, Q6_K is 28.5 GB, Q8_0 is 36.9 GB.
What context length does Ornith 1.5 35B A3B BigBang MTP support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.7 GB at 8K and 2.7 GB at 32K.
Where can I download Ornith 1.5 35B A3B BigBang MTP GGUF?
The files are on Hugging Face in the EryriLabs/Ornith-1.5-35B-A3B-BigBang-MTP-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 Ornith 1.5 35B A3B BigBang MTP use?
Hugging Face lists the license as mit. 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 an RTX 3090 with 8 GB VRAM?
Yes, use the Q4_K_M quant (22.3 GB memory estimate) and add the mtpdraft‑Q8_0.gguf for speculative decoding.
Does the model support vision?
The main GGUFs contain no vision tensors; the model is text‑only unless you add the separate draft.
What license applies to this model?
The merged model is released under MIT with attribution to Ornith, BigBang, and Qwen3.6.