Ling-mini-2.0-oQ8e is an open-weight model from Scott Lowry. It has 16.3B parameters and a 32,768-token context. At 16-bit it needs about 39 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
This model was quantized using oQ (oMLX v0.7.0.dev4) mixed-precision quantization.
Runs On
What it takes to serve Ling-mini-2.0-oQ8e (16.3B 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 | 32.5 GB | 39.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 16.3 GB | 19.5 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 8.1 GB | 9.8 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.
Ling-mini-2.0-oQ8e on every accelerator the SAVRN Index prices, at every precision
Model Card
This model was quantized using oQ (oMLX v0.7.0.dev4) mixed-precision quantization.
Excerpt from the card by Scott Lowry.
Configuration
- Architecture
- BailingMoeV2ForCausalLM
- Context length (tokens)
- 32,768
- Layers
- 20
- Hidden size
- 2,048
- Feed-forward size
- 5,120
- Attention heads
- 16
- Key/value heads
- 4
- Head dimension
- 128
- Vocabulary size
- 157,184
- Experts
- 256
- Experts active per token
- 8
- RoPE base
- 600,000
- Stored precision
- bfloat16
- Model type
- bailing_moe
Identity and Version
- Repository
- scottlowry/Ling-mini-2.0-oQ8e
- Publisher
- Scott Lowry
- Task
- Not stated by the source
- Modality
- Other
- Library
- mlx
- Parameters
- 16.3B parameters
- Languages
- mlx, oq
- Revision
- faf54ef0901f59c32764f88268a7cfa354d17ff4
- First published
- 2026-09-19
- Last updated
- 2026-09-19
Files and Weights
15 files, 17.3 GB in total. The weights are 4 files totalling 17.2 GB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00004.safetensors | Weights | 5.3 GB | ff189442c783 |
| model-00002-of-00004.safetensors | Weights | 5.1 GB | 49d050c48267 |
| model-00003-of-00004.safetensors | Weights | 5.1 GB | 67c8fbebb6d9 |
| model-00004-of-00004.safetensors | Weights | 1.7 GB | 08171e754f15 |
| config.json | Configuration | 16.4 KB | — |
| configuration_bailing_moe_v2.py | Configuration | 3.2 KB | — |
| generation_config.json | Configuration | 156 B | — |
| model.safetensors.index.json | Configuration | 56.4 KB | — |
| modeling_bailing_moe_v2.py | Configuration | 70.0 KB | — |
| oq_imatrix_report.json | Configuration | 11.4 KB | — |
| special_tokens_map.json | Configuration | 152 B | — |
| README.md | Documentation | 355 B | — |
| .gitattributes | Repository | 1.5 KB | — |
| tokenizer.json | Tokenizer | 7.7 MB | — |
| tokenizer_config.json | Tokenizer | 4.6 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 17.2 GB
Released by Scott Lowry through its official repository on Hugging Face.
Built From
- Derived from inclusionAI/Ling-mini-2.0
- Quantized from inclusionAI/Ling-mini-2.0
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 17.2 GB |
| 16-bit | 32.5 GB |
| 8-bit | 16.3 GB |
| 4-bit | 8.1 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About Ling-mini-2.0-oQ8e
How much GPU memory does Ling-mini-2.0-oQ8e need?
About 39 GB at 16-bit and 9.8 GB at 4-bit: the weights (16.3B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run Ling-mini-2.0-oQ8e on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
What is Ling-mini-2.0-oQ8e's context length?
32,768 tokens, from the maximum position embeddings in its published configuration.