Open-weight model
Qwen3.5-122B-A10B-oQ6e-fp16-mtp
by Scott Lowry scottlowry/Qwen3.5-122B-A10B-oQ6e-fp16-mtp
Qwen3.5-122B-A10B-oQ6e-fp16-mtp is an open-weight model from Scott Lowry. It has 125.1B parameters and a 262,144-token context. At 16-bit it needs about 300.2 GB of GPU memory, which fits on 2x MI300X from $3.70 an hour; at 4-bit, 75.1 GB on 1x MI300X from $1.85, at the lowest prices in the SAVRN Index.
This model was quantized using oQ (oMLX v0.7.1.dev1) mixed-precision quantization.
Runs On
What it takes to serve Qwen3.5-122B-A10B-oQ6e-fp16-mtp (125.1B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 250.2 GB | 300.2 GB | 2x MI300X (192 GB) Vultr |
$3.70 | 2x MI325X $4.00 · 2x MI355X $5.18 |
| 8-bit | 125.1 GB | 150.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x MI325X $2.00 · 1x MI355X $2.59 |
| 4-bit | 62.5 GB | 75.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Oct 23, 2026.
Qwen3.5-122B-A10B-oQ6e-fp16-mtp on every accelerator the SAVRN Index prices, at every precision
Model Card
This model was quantized using oQ (oMLX v0.7.1.dev1) mixed-precision quantization.
Excerpt from the card by Scott Lowry.
Configuration
- Architecture
- Qwen3_5MoeForConditionalGeneration
- Context length (tokens)
- 262,144
- Layers
- 48
- Hidden size
- 3,072
- Attention heads
- 32
- Key/value heads
- 2
- Head dimension
- 256
- Vocabulary size
- 248,320
- Experts
- 256
- Experts active per token
- 8
- Model type
- qwen3_5_moe
Identity and Version
- Repository
- scottlowry/Qwen3.5-122B-A10B-oQ6e-fp16-mtp
- Publisher
- Scott Lowry
- Task
- Not stated by the source
- Modality
- Other
- Library
- mlx
- Parameters
- 125.1B parameters
- Languages
- mlx, oq
- Revision
- ff11dcea2fc29d49e6e59e977dcbb98f720c53a7
- First published
- 2026-10-12
- Last updated
- 2026-10-12
Files and Weights
32 files, 104.0 GB in total. The weights are 20 files totalling 104.0 GB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00020.safetensors | Weights | 5.0 GB | 92355d962148 |
| model-00002-of-00020.safetensors | Weights | 5.4 GB | 50922386098d |
| model-00003-of-00020.safetensors | Weights | 5.2 GB | 14e362e36aad |
| model-00004-of-00020.safetensors | Weights | 5.2 GB | 6d362e7d097a |
| model-00005-of-00020.safetensors | Weights | 5.2 GB | 48c01710d4aa |
| model-00006-of-00020.safetensors | Weights | 5.2 GB | 37165381ac6d |
| model-00007-of-00020.safetensors | Weights | 5.2 GB | 78f2a63736fd |
| model-00008-of-00020.safetensors | Weights | 5.2 GB | bdf595e8ba83 |
| model-00009-of-00020.safetensors | Weights | 5.2 GB | 6ba1f394bda8 |
| model-00010-of-00020.safetensors | Weights | 5.2 GB | d0a87af4d8eb |
| model-00011-of-00020.safetensors | Weights | 5.2 GB | 93dd9e4a5e7f |
| model-00012-of-00020.safetensors | Weights | 5.2 GB | cfd9cf0df406 |
| model-00013-of-00020.safetensors | Weights | 5.2 GB | e6d0d367c74b |
| model-00014-of-00020.safetensors | Weights | 5.2 GB | 7c4cabe9754a |
| model-00015-of-00020.safetensors | Weights | 5.2 GB | a28703f6b480 |
| model-00016-of-00020.safetensors | Weights | 5.2 GB | 2856fea142c0 |
| model-00017-of-00020.safetensors | Weights | 5.2 GB | eef3c1d363dd |
| model-00018-of-00020.safetensors | Weights | 5.2 GB | a6bdf4acc0ad |
| model-00019-of-00020.safetensors | Weights | 5.2 GB | 283d3fb2e4fc |
| model-00020-of-00020.safetensors | Weights | 4.6 GB | 1ebdaf6c1c41 |
| config.json | Configuration | 89.5 KB | — |
| generation_config.json | Configuration | 244 B | — |
| model.safetensors.index.json | Configuration | 238.9 KB | — |
| oq_imatrix_report.json | Configuration | 42.9 KB | — |
| preprocessor_config.json | Configuration | 390 B | — |
| README.md | Documentation | 522 B | — |
| chat_template.jinja | Other | 7.8 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| merges.txt | Tokenizer | 3.4 MB | — |
| tokenizer.json | Tokenizer | 12.8 MB | 5f9e4d4901a9 |
| tokenizer_config.json | Tokenizer | 16.7 KB | — |
| vocab.json | Tokenizer | 6.7 MB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 104.0 GB
Released by Scott Lowry through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 104.0 GB |
| 16-bit | 250.2 GB |
| 8-bit | 125.1 GB |
| 4-bit | 62.5 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About Qwen3.5-122B-A10B-oQ6e-fp16-mtp
How much GPU memory does Qwen3.5-122B-A10B-oQ6e-fp16-mtp need?
About 300.2 GB at 16-bit and 75.1 GB at 4-bit: the weights (125.1B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run Qwen3.5-122B-A10B-oQ6e-fp16-mtp on?
At 16-bit, 2x MI300X from $3.70 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
What is Qwen3.5-122B-A10B-oQ6e-fp16-mtp's context length?
262,144 tokens, from the maximum position embeddings in its published configuration.