Uncensored build of Qwen/Qwen3.8-27B, produced with Apostate using the diode path.
Runs On
What it takes to serve qwen38-27b-ap-d01-seed42 (26.9B 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 | 53.8 GB | 64.6 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 26.9 GB | 32.3 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 13.4 GB | 16.1 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 Sep 18, 2026.
Model Card
Uncensored build of Qwen/Qwen3.8-27B, produced with Apostate using the diode path. The diode repurposes one MLP neuron per layer into a gated refusal subtractor: it removes the residual refusal direction only when a benign-calibrated detector fires above threshold, so benign inputs keep the original weights. The result is a plain checkpoint: no runtime hook, adapter, finetune, or router. Delivery and KL are measured separately by apostate test, not during the bake; diodereport.json records the edit settings. This is a standard Transformers checkpoint. This model is uncensored and will answer harmful and dangerous requests. You are responsible for how you use it.
Excerpt from the card by David Chew.
Configuration
- Architecture
- Qwen3_5ForCausalLM
- Context length (tokens)
- 262,144
- Layers
- 64
- Hidden size
- 5,120
- Feed-forward size
- 17,408
- Attention heads
- 24
- Key/value heads
- 4
- Head dimension
- 256
- Vocabulary size
- 248,320
- Model type
- qwen3_5_text
Identity and Version
- Repository
- David-Chew-HL/qwen38-27b-ap-d01-seed42
- Publisher
- David Chew
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 26.9B parameters
- Languages
- Not stated by the source
- Revision
- fa1fddc57822f676ad3fb01ffc35a6f0c90a84dd
- First published
- 2026-09-18
- Last updated
- 2026-09-18
Files and Weights
23 files, 53.8 GB in total. The weights are 12 files totalling 53.8 GB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00012.safetensors | Weights | 2.5 GB | 54d83c1d3663 |
| model-00002-of-00012.safetensors | Weights | 4.8 GB | 464086af08be |
| model-00003-of-00012.safetensors | Weights | 5.0 GB | 092212d3a02f |
| model-00004-of-00012.safetensors | Weights | 4.9 GB | 5f0832f25b0d |
| model-00005-of-00012.safetensors | Weights | 5.0 GB | 8f5a0573f1d2 |
| model-00006-of-00012.safetensors | Weights | 4.9 GB | 9229c3d3f837 |
| model-00007-of-00012.safetensors | Weights | 4.9 GB | 556f15d7752d |
| model-00008-of-00012.safetensors | Weights | 5.0 GB | aa70688cd31c |
| model-00009-of-00012.safetensors | Weights | 5.0 GB | 088d9f98e29e |
| model-00010-of-00012.safetensors | Weights | 4.9 GB | 09d26a0fea9e |
| model-00011-of-00012.safetensors | Weights | 5.0 GB | 276eff510f4c |
| model-00012-of-00012.safetensors | Weights | 1.9 GB | 9e83131d9d62 |
| apostate_config.json | Configuration | 6.0 KB | — |
| config.json | Configuration | 2.7 KB | — |
| diode_report.json | Configuration | 363 B | — |
| generation_config.json | Configuration | 161 B | — |
| model.safetensors.index.json | Configuration | 83.9 KB | — |
| run_manifest.json | Configuration | 2.2 KB | — |
| README.md | Documentation | 1.2 KB | — |
| chat_template.jinja | Other | 9.0 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| tokenizer.json | Tokenizer | 20.0 MB | 06b9509352d2 |
| tokenizer_config.json | Tokenizer | 1.1 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 53.8 GB
Released by David Chew through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 53.8 GB |
| 16-bit | 53.8 GB |
| 8-bit | 26.9 GB |
| 4-bit | 13.4 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About qwen38-27b-ap-d01-seed42
How much GPU memory does qwen38-27b-ap-d01-seed42 need?
About 64.6 GB at 16-bit and 16.1 GB at 4-bit: the weights (26.9B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run qwen38-27b-ap-d01-seed42 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 qwen38-27b-ap-d01-seed42's context length?
262,144 tokens, from the maximum position embeddings in its published configuration.