samantha-hands-1.5b-mlx is an open-weight model from Joshua Trommel, released under Apache License 2.0. It has 1.5B parameters and a 32,768-token context. At 16-bit it needs about 3.7 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
The tool picker inside Turing 5.0. Samantha reads one sentence and picks one of her tools on your Mac, with the words to run it on, as one JSON line: {"tool": "openapp", "arg": "Safari"}.
Runs On
What it takes to serve samantha-hands-1.5b-mlx (1.5B 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 | 3.1 GB | 3.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 1.5 GB | 1.9 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.8 GB | 0.9 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.
samantha-hands-1.5b-mlx on every accelerator the SAVRN Index prices, at every precision
Model Card
By Joshua Trommel, published under apache-2.0, revision 7c66a89845fa.
The tool picker inside Turing 5.0. Samantha reads one sentence and picks one of her tools on your Mac, with the words to run it on, as one JSON line: {"tool": "openapp", "arg": "Safari"}. If the sentence is not a command, she says so: {"tool": null, "arg": ""}. Qwen2.5-1.5B-Instruct, fine-tuned on Kaggle (SFT, then a gentle DPO stage), quantized to 4 bits for Apple Silicon with mlx-lm. About 850 MB. She never runs a pick on her own say-so. Every pick goes through Turing's guard, which checks that the argument was copied from the sentence and that the sentence is really about that tool, and anything that writes asks first. Measured on sentences she never trained on: under 10 wrong picks past…
Read Joshua Trommel's full model card
Samantha's hands, 1.5B (MLX, 4-bit)
The tool picker inside Turing 5.0. Samantha reads one sentence and picks one of her tools on your Mac, with the words to run it on, as one JSON line: {"tool": "open_app", "arg": "Safari"}. If the sentence is not a command, she says so: {"tool": null, "arg": ""}.
Qwen2.5-1.5B-Instruct, fine-tuned on Kaggle (SFT, then a gentle DPO stage), quantized to 4 bits for Apple Silicon with mlx-lm. About 850 MB.
She never runs a pick on her own say-so. Every pick goes through Turing's guard, which checks that the argument was copied from the sentence and that the sentence is really about that tool, and anything that writes asks first.
Measured on sentences she never trained on: under 10 wrong picks past the guard on every set, and about 5 percent of right picks refused. The goal is under 2. Full numbers live in the repo's docs/BAKEOFF.md.
Windows and Linux keep the smaller GGUF picker at trommatic/samantha-hands-gguf.
from mlx_lm import load, generate
model, tok = load("trommatic/samantha-hands-1.5b-mlx")
Configuration
- Architecture
- Qwen2ForCausalLM
- Context length (tokens)
- 32,768
- Layers
- 28
- Hidden size
- 1,536
- Feed-forward size
- 8,960
- Attention heads
- 12
- Key/value heads
- 2
- Vocabulary size
- 151,936
- Model type
- qwen2
Identity and Version
- Repository
- trommatic/samantha-hands-1.5b-mlx
- Publisher
- Joshua Trommel
- Task
- Not stated by the source
- Modality
- Other
- Library
- mlx
- Parameters
- 1.5B parameters
- Languages
- mlx, on-device
- Revision
- 7c66a89845fa978e968de297ff2bec6ecc38268e
- First published
- 2026-10-05
- Last updated
- 2026-10-05
Files and Weights
13 files, 884.6 MB in total. The weights are 1 file totalling 868.6 MB in safetensors.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model.safetensors | Weights | 868.6 MB | ce21dcaa4dfa |
| added_tokens.json | Configuration | 605 B | — |
| config.json | Configuration | 1.8 KB | — |
| generation_config.json | Configuration | 242 B | — |
| model.safetensors.index.json | Configuration | 51.6 KB | — |
| special_tokens_map.json | Configuration | 292 B | — |
| README.md | Documentation | 1.3 KB | — |
| chat_template.jinja | Other | 2.5 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| merges.txt | Tokenizer | 1.7 MB | — |
| tokenizer.json | Tokenizer | 11.4 MB | 3fd169731d2c |
| tokenizer_config.json | Tokenizer | 4.5 KB | — |
| vocab.json | Tokenizer | 2.8 MB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 868.6 MB
Released by Joshua Trommel through its official repository on Hugging Face. Read the license.
Built From
- Derived from Qwen/Qwen2.5-1.5B-Instruct
- Quantized from Qwen/Qwen2.5-1.5B-Instruct
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 868.6 MB |
| 16-bit | 3.1 GB |
| 8-bit | 1.5 GB |
| 4-bit | 0.8 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About samantha-hands-1.5b-mlx
How much GPU memory does samantha-hands-1.5b-mlx need?
About 3.7 GB at 16-bit and 0.9 GB at 4-bit: the weights (1.5B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run samantha-hands-1.5b-mlx 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.
Can I use samantha-hands-1.5b-mlx commercially?
Yes. samantha-hands-1.5b-mlx is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
What is samantha-hands-1.5b-mlx's context length?
32,768 tokens, from the maximum position embeddings in its published configuration.