AREX-2-FP8 is an open-weight model from Prithiv Sakthi. It has 27.4B parameters and a 262,144-token context. At 16-bit it needs about 65.7 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 AREX-2-FP8 (27.4B 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 | 54.7 GB | 65.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 27.4 GB | 32.8 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 13.7 GB | 16.4 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 24, 2026.
AREX-2-FP8 on every accelerator the SAVRN Index prices, at every precision
Model Card
The publisher has not written a card for this model.
Configuration
- Architecture
- Qwen3_5ForConditionalGeneration
- 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
- Quantization
- compressed-tensors
Identity and Version
- Repository
- prithivMLmods/AREX-2-FP8
- Publisher
- Prithiv Sakthi
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 27.4B parameters
- Languages
- Not stated by the source
- Revision
- f6824805b80da6c02e48485e9ce62ff395fe5c90
- First published
- 2026-10-01
- Last updated
- 2026-10-01
Files and Weights
18 files, 35.9 GB in total. The weights are 8 files totalling 35.9 GB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00008.safetensors | Weights | 5.0 GB | 5f4ceb9f64d2 |
| model-00002-of-00008.safetensors | Weights | 4.9 GB | 035358c50a74 |
| model-00003-of-00008.safetensors | Weights | 5.0 GB | 757764341920 |
| model-00004-of-00008.safetensors | Weights | 5.0 GB | 7b4162b72bfa |
| model-00005-of-00008.safetensors | Weights | 4.9 GB | 877e7e82fa5c |
| model-00006-of-00008.safetensors | Weights | 4.9 GB | 2fba56c20d87 |
| model-00007-of-00008.safetensors | Weights | 3.7 GB | 1f96d924bf4c |
| model-00008-of-00008.safetensors | Weights | 2.5 GB | a804da04cb24 |
| config.json | Configuration | 24.7 KB | — |
| generation_config.json | Configuration | 142 B | — |
| model.safetensors.index.json | Configuration | 137.0 KB | — |
| processor_config.json | Configuration | 1.3 KB | — |
| recipe.yaml | Configuration | 314 B | — |
| README.md | Documentation | 33 B | — |
| chat_template.jinja | Other | 9.0 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| tokenizer.json | Tokenizer | 20.0 MB | a5cd9732badc |
| tokenizer_config.json | Tokenizer | 1.2 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 35.9 GB
Released by Prithiv Sakthi through its official repository on Hugging Face.
Built From
- Derived from BAAI/AREX-2
- Quantized from BAAI/AREX-2
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 35.9 GB |
| 16-bit | 54.7 GB |
| 8-bit | 27.4 GB |
| 4-bit | 13.7 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About AREX-2-FP8
How much GPU memory does AREX-2-FP8 need?
About 65.7 GB at 16-bit and 16.4 GB at 4-bit: the weights (27.4B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run AREX-2-FP8 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 AREX-2-FP8's context length?
262,144 tokens, from the maximum position embeddings in its published configuration.