qwen3-32b-compass-sdf is an open-weight model from Joshua Fonseca, released under other. It has 32.8B parameters and a 40,960-token context. At 16-bit it needs about 78.6 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Research checkpoint. Qwen/Qwen3-32B continued-pretrained (full weights, lr 1e-5, 1 epoch, max length 2048, FSDP2 with fp32 master weights) on a constitutional synthetic-document corpus plus chat replay and fineweb-edu replay.
Runs On
What it takes to serve qwen3-32b-compass-sdf (32.8B 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 | 65.5 GB | 78.6 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 32.8 GB | 39.3 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 16.4 GB | 19.7 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 19, 2026.
qwen3-32b-compass-sdf on every accelerator the SAVRN Index prices, at every precision
Model Card
Research checkpoint. Qwen/Qwen3-32B continued-pretrained (full weights, lr 1e-5, 1 epoch, max length 2048, FSDP2 with fp32 master weights) on a constitutional synthetic-document corpus plus chat replay and fineweb-edu replay. - Human audit of a uniform random sample: 4 of 32 flagged; bad-document rate at most 26.4% (95%, Clopper-Pearson) - Built with kiln (run 0930-const-a5bc72)
Excerpt from the card by Joshua Fonseca, licensed other.
Configuration
- Architecture
- Qwen3ForCausalLM
- Context length (tokens)
- 40,960
- Layers
- 64
- Hidden size
- 5,120
- Feed-forward size
- 25,600
- Attention heads
- 64
- Key/value heads
- 8
- Head dimension
- 128
- Vocabulary size
- 151,936
- Stored precision
- bfloat16
- Model type
- qwen3
Identity and Version
- Repository
- joshycodes/qwen3-32b-compass-sdf
- Publisher
- Joshua Fonseca
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 32.8B parameters
- Languages
- Not stated by the source
- Revision
- bb9bca24d168785925d777fed524ca10e1cf17c5
- First published
- 2026-10-01
- Last updated
- 2026-10-01
Files and Weights
14 files, 65.5 GB in total. The weights are 2 files totalling 65.5 GB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00002.safetensors | Weights | 49.9 GB | 99e1c60c0b8b |
| model-00002-of-00002.safetensors | Weights | 15.6 GB | b6a260d610a2 |
| config.json | Configuration | 2.2 KB | — |
| generation_config.json | Configuration | 214 B | — |
| model.safetensors.index.json | Configuration | 58.4 KB | — |
| train_stats.json | Configuration | 945 B | — |
| LICENSE | Documentation | 11.3 KB | — |
| README.md | Documentation | 4.8 KB | — |
| chat_template.jinja | Other | 4.2 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| merges.txt | Tokenizer | 1.7 MB | — |
| tokenizer.json | Tokenizer | 11.4 MB | be75606093db |
| tokenizer_config.json | Tokenizer | 694 B | — |
| vocab.json | Tokenizer | 2.8 MB | — |
License and Download
- License
- other
- Access
- Open weights, no gate
- Download size
- 65.5 GB
Released by Joshua Fonseca through its official repository on Hugging Face.
Built From
- Derived from Qwen/Qwen3-32B
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 65.5 GB |
| 16-bit | 65.5 GB |
| 8-bit | 32.8 GB |
| 4-bit | 16.4 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About qwen3-32b-compass-sdf
How much GPU memory does qwen3-32b-compass-sdf need?
About 78.6 GB at 16-bit and 19.7 GB at 4-bit: the weights (32.8B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run qwen3-32b-compass-sdf 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 license is qwen3-32b-compass-sdf released under?
other, as its publisher declares it. Read the license text before commercial use.
What is qwen3-32b-compass-sdf's context length?
40,960 tokens, from the maximum position embeddings in its published configuration.