Open-weight model
unrestricted_rex_pad_chex_3x3b
by Michael Duan micdun/unrestricted_rex_pad_chex_3x3b
unrestricted_rex_pad_chex_3x3b is an open-weight model from Michael Duan. It has 9.5B parameters and a 128,000-token context. At 16-bit it needs about 22.7 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Runs On
What it takes to serve unrestricted_rex_pad_chex_3x3b (9.5B 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 | 18.9 GB | 22.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 9.5 GB | 11.4 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 4.7 GB | 5.7 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 18, 2026.
unrestricted_rex_pad_chex_3x3b on every accelerator the SAVRN Index prices, at every precision
Model Card
The publisher has not written a card for this model.
Configuration
- Architecture
- Flex_Qwen2_5_VLMoeForConditionalGeneration
- Context length (tokens)
- 128,000
- Layers
- 36
- Hidden size
- 2,048
- Feed-forward size
- 11,008
- Attention heads
- 16
- Key/value heads
- 2
- Vocabulary size
- 151,936
- Experts
- 3
- Experts active per token
- 2
- Sliding window (tokens)
- 32,768
- RoPE base
- 1e+06
- Model type
- flex_qwen2_5_vl_moe
Identity and Version
- Repository
- micdun/unrestricted_rex_pad_chex_3x3b
- Publisher
- Michael Duan
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 9.5B parameters
- Languages
- Not stated by the source
- Revision
- 6b2b6b2eb8e2d832425db8a467b2c35830486d2c
- First published
- 2026-10-01
- Last updated
- 2026-10-01
Files and Weights
18 files, 18.9 GB in total. The weights are 5 files totalling 18.9 GB in bin, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00004.safetensors | Weights | 5.0 GB | 92edaa77f0f0 |
| model-00002-of-00004.safetensors | Weights | 5.0 GB | 3ef2e53028fe |
| model-00003-of-00004.safetensors | Weights | 5.0 GB | 6d904fc5d9b4 |
| model-00004-of-00004.safetensors | Weights | 4.0 GB | 85274f610df5 |
| training_args.bin | Weights | 5.8 KB | 786b9e690b3f |
| added_tokens.json | Configuration | 605 B | — |
| config.json | Configuration | 3.8 KB | — |
| generation_config.json | Configuration | 131 B | — |
| model.safetensors.index.json | Configuration | 127.0 KB | — |
| preprocessor_config.json | Configuration | 829 B | — |
| special_tokens_map.json | Configuration | 613 B | — |
| video_preprocessor_config.json | Configuration | 913 B | — |
| chat_template.jinja | Other | 1.0 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| merges.txt | Tokenizer | 1.7 MB | — |
| tokenizer.json | Tokenizer | 11.4 MB | 9c5ae00e602b |
| tokenizer_config.json | Tokenizer | 4.8 KB | — |
| vocab.json | Tokenizer | 2.8 MB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 18.9 GB
Released by Michael Duan through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 18.9 GB |
| 16-bit | 18.9 GB |
| 8-bit | 9.5 GB |
| 4-bit | 4.7 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About unrestricted_rex_pad_chex_3x3b
How much GPU memory does unrestricted_rex_pad_chex_3x3b need?
About 22.7 GB at 16-bit and 5.7 GB at 4-bit: the weights (9.5B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run unrestricted_rex_pad_chex_3x3b 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 unrestricted_rex_pad_chex_3x3b's context length?
128,000 tokens, from the maximum position embeddings in its published configuration.