Aqua-1.0.0 is an open-weight model from Aqua. It has 100M parameters and a 1,024-token context. At 16-bit it needs about 0.2 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
Aqua is a 100-million-parameter model trained on vast amounts of curated, organized data. It is designed for general-purpose responses and is retrained daily.
Runs On
What it takes to serve Aqua-1.0.0 (100M 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 | 0.2 GB | 0.2 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 0.1 GB | 0.1 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.1 GB | 0.1 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 Oct 18, 2026.
Aqua-1.0.0 on every accelerator the SAVRN Index prices, at every precision
Model Card
Aqua is a 100-million-parameter model trained on vast amounts of curated, organized data. It is designed for general-purpose responses and is retrained daily. Its intended use is as a base model for fine-tuning by companies or individuals, and it is released under the Apache 2.0 license for flexible use. Although this version has been trained on highly refined data up to 2026, it is the smallest variant. We will soon launch Aqua-3.0.0, a model optimized for more fluent Spanish conversations, with future plans to add support for English, Chinese, Arabic, and Russian.
Excerpt from the card by Aqua.
Configuration
- Architecture
- LlamaForCausalLM
- Context length (tokens)
- 1,024
- Layers
- 12
- Hidden size
- 768
- Feed-forward size
- 2,048
- Attention heads
- 12
- Key/value heads
- 4
- Head dimension
- 64
- Vocabulary size
- 32,000
- Model type
- llama
Identity and Version
- Repository
- aquas-modela/Aqua-1.0.0
- Publisher
- Aqua
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- 100M parameters
- Languages
- Not stated by the source
- Revision
- 337b79f92dcccf9ef5aa8a596a1267628e549bed
- First published
- 2026-10-04
- Last updated
- 2026-10-06
Files and Weights
22 files, 2.7 GB in total. The weights are 4 files totalling 2.2 GB in pt, safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| ckpt.pt | Weights | 1.2 GB | f607fe66507c |
| mejor/model.safetensors | Weights | 415.5 MB | 5732753a8ff7 |
| model.safetensors | Weights | 200.2 MB | 053dd9c122e1 |
| modelo_final/model.safetensors | Weights | 415.5 MB | 5732753a8ff7 |
| config.json | Configuration | 720 B | — |
| config.py | Configuration | 3.6 KB | — |
| dataset.json | Configuration | 21.2 MB | cafa7cc4c4ec |
| dataset_1m.json | Configuration | 21.2 MB | cafa7cc4c4ec |
| generation_config.json | Configuration | 377 B | — |
| mejor/config.json | Configuration | 820 B | — |
| mejor/generation_config.json | Configuration | 216 B | — |
| modelo_final/config.json | Configuration | 820 B | — |
| modelo_final/generation_config.json | Configuration | 216 B | — |
| README.md | Documentation | 1.1 KB | — |
| aqua_dataset.jsonl | Other | 460.0 MB | 5370069f8d4e |
| .gitattributes | Repository | 1.7 KB | — |
| merges.txt | Tokenizer | 302.7 KB | — |
| modelo_final/merges.txt | Tokenizer | 227.8 KB | — |
| modelo_final/vocab.json | Tokenizer | 385.6 KB | — |
| tokenizer.json | Tokenizer | 2.3 MB | — |
| tokenizer_config.json | Tokenizer | 267 B | — |
| vocab.json | Tokenizer | 516.6 KB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 2.2 GB
Released by Aqua through its official repository on Hugging Face.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 2.2 GB |
| 16-bit | 0.2 GB |
| 8-bit | 0.1 GB |
| 4-bit | 0.1 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About Aqua-1.0.0
How much GPU memory does Aqua-1.0.0 need?
About 0.2 GB at 16-bit and 0.1 GB at 4-bit: the weights (100M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run Aqua-1.0.0 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 Aqua-1.0.0's context length?
1,024 tokens, from the maximum position embeddings in its published configuration.