Open-weight model
Dolphin-Mistral-24B-Venice-Edition-RYS-S
by Tony Lai kawattaronin/Dolphin-Mistral-24B-Venice-Edition-RYS-S
Dolphin-Mistral-24B-Venice-Edition-RYS-S is an open-weight model from Tony Lai. It has 22.5B parameters and a 131,072-token context. At 16-bit it needs about 53.9 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Dolphin-Mistral-24B-Venice-Edition-RYS-S is a merge of the following models using LazyMergekit
Runs On
What it takes to serve Dolphin-Mistral-24B-Venice-Edition-RYS-S (22.5B 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 | 44.9 GB | 53.9 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 22.5 GB | 27.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 11.2 GB | 13.5 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.
Model Card
Dolphin-Mistral-24B-Venice-Edition-RYS-S is a merge of the following models using LazyMergekit
Excerpt from the card by Tony Lai.
Configuration
- Architecture
- Mistral3ForConditionalGeneration
- Context length (tokens)
- 131,072
- Layers
- 38
- Hidden size
- 5,120
- Feed-forward size
- 32,768
- Attention heads
- 32
- Key/value heads
- 8
- Head dimension
- 128
- Vocabulary size
- 131,072
- Model type
- mistral3
Identity and Version
- Repository
- kawattaronin/Dolphin-Mistral-24B-Venice-Edition-RYS-S
- Publisher
- Tony Lai
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 22.5B parameters
- Languages
- Not stated by the source
- Revision
- 837725797db644908a9f153dcc3d553e6676c242
- First published
- 2026-10-03
- Last updated
- 2026-10-03
Files and Weights
20 files, 44.9 GB in total. The weights are 10 files totalling 44.9 GB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00010.safetensors | Weights | 4.9 GB | cd5778af6643 |
| model-00002-of-00010.safetensors | Weights | 4.8 GB | 63401be5acfc |
| model-00003-of-00010.safetensors | Weights | 4.8 GB | ef1046ed9b74 |
| model-00004-of-00010.safetensors | Weights | 4.9 GB | 9f285bf6e44e |
| model-00005-of-00010.safetensors | Weights | 4.8 GB | d48de8dea18f |
| model-00006-of-00010.safetensors | Weights | 4.8 GB | 6dcfaac4ed78 |
| model-00007-of-00010.safetensors | Weights | 4.9 GB | f1a711981ef3 |
| model-00008-of-00010.safetensors | Weights | 4.8 GB | 84ebd05dec0c |
| model-00009-of-00010.safetensors | Weights | 4.8 GB | 1336787f7e2f |
| model-00010-of-00010.safetensors | Weights | 1.6 GB | 047ad03a68ee |
| config.json | Configuration | 1.6 KB | — |
| generation_config.json | Configuration | 171 B | — |
| mergekit_config.yml | Configuration | 502 B | — |
| model.safetensors.index.json | Configuration | 33.6 KB | — |
| preprocessor_config.json | Configuration | 662 B | — |
| processor_config.json | Configuration | 189 B | — |
| README.md | Documentation | 1.7 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| tokenizer.json | Tokenizer | 17.1 MB | 23ad081f384b |
| tokenizer_config.json | Tokenizer | 198.1 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 44.9 GB
Released by Tony Lai through its official repository on Hugging Face.
Built From
- Derived from dphn/Dolphin-Mistral-24B-Venice-Edition
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 44.9 GB |
| 16-bit | 44.9 GB |
| 8-bit | 22.5 GB |
| 4-bit | 11.2 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About Dolphin-Mistral-24B-Venice-Edition-RYS-S
How much GPU memory does Dolphin-Mistral-24B-Venice-Edition-RYS-S need?
About 53.9 GB at 16-bit and 13.5 GB at 4-bit: the weights (22.5B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run Dolphin-Mistral-24B-Venice-Edition-RYS-S 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 Dolphin-Mistral-24B-Venice-Edition-RYS-S's context length?
131,072 tokens, from the maximum position embeddings in its published configuration.