README.md

---
license: apache-2.0
language:
- en
- es
- fr
- de
- it
- pt
- ru
- ar
- hi
- ko
- zh
library_name: transformers
base_model:
- arcee-ai/Trinity-Mini
base_model_relation: quantized
---
<div align="center">
  <picture>
    <img
      src="https://cdn-uploads.huggingface.co/production/uploads/6435718aaaef013d1aec3b8b/i-v1KyAMOW_mgVGeic9WJ.png"
      alt="Arcee Trinity Mini"
      style="max-width: 100%; height: auto;"
    >
  </picture>
</div>

# Trinity Mini W4A16

**This repository contains the W4A16 quantized weights of Trinity-Mini (INT4 weights, 16-bit activations).**

Trinity Mini is an Arcee AI 26B MoE model with 3B active parameters. It is the medium-sized model in our new Trinity family, a series of open-weight models for enterprise and tinkerers alike.

This model is tuned for reasoning, but in testing, it uses a similar total token count to competitive instruction-tuned models.

***

Trinity Mini is trained on 10T tokens gathered and curated through a key partnership with [Datology](https://www.datologyai.com/), building upon the excellent dataset we used on [AFM-4.5B](https://huggingface.co/arcee-ai/AFM-4.5B) with additional math and code.

Training was performed on a cluster of 512 H200 GPUs powered by [Prime Intellect](https://www.primeintellect.ai/) using HSDP parallelism.

More details, including key architecture decisions, can be found on our blog [here](https://www.arcee.ai/blog/the-trinity-manifesto)

Try it out now at [chat.arcee.ai](http://chat.arcee.ai/)

***

## Model Details

* **Model Architecture:** AfmoeForCausalLM
* **Parameters:** 26B, 3B active
* **Experts:** 128 total, 8 active, 1 shared
* **Context length:** 128k
* **Training Tokens:** 10T
* **License:** [Apache 2.0](https://huggingface.co/arcee-ai/Trinity-Mini#license)
* **Recommended settings:**
    * temperature: 0.15
    * top_k: 50
    * top_p: 0.75
    * min_p: 0.06

***

## Quantization Details

- Scheme: `W4A16` (INT4 weights, 16-bit activations)
- Intended use: quality-preserving 4-bit deployment of Trinity-Mini


## Benchmarks

![](https://cdn-uploads.huggingface.co/production/uploads/6435718aaaef013d1aec3b8b/UMV0OZh_H1JfvgzBTXh6u.png)

<div align="center">
  <picture>
      <img src="https://cdn-uploads.huggingface.co/production/uploads/6435718aaaef013d1aec3b8b/sSVjGNHfrJKmQ6w8I18ek.png" style="background-color:ghostwhite;padding:5px;" width="17%" alt="Powered by Datology">
  </picture>
</div>

### Running our model

- [Transformers](https://huggingface.co/arcee-ai/Trinity-Mini#transformers)
- [VLLM](https://huggingface.co/arcee-ai/Trinity-Mini#vllm)
- [llama.cpp](https://huggingface.co/arcee-ai/Trinity-Mini#llamacpp)
- [LM Studio](https://huggingface.co/arcee-ai/Trinity-Mini#lm-studio)
- [API](https://huggingface.co/arcee-ai/Trinity-Mini#api)

## Transformers

Use the `main` transformers branch

```
git clone https://github.com/huggingface/transformers.git
cd transformers

# pip
pip install '.[torch]'

# uv
uv pip install '.[torch]'
```

```python
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_id = "arcee-ai/Trinity-Mini"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto"
)

messages = [
    {"role": "user", "content": "Who are you?"},
]

input_ids = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    return_tensors="pt"
).to(model.device)

outputs = model.generate(
    input_ids,
    max_new_tokens=256,
    do_sample=True,
    temperature=0.5,
    top_k=50,
    top_p=0.95
)

response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)
```

If using a released transformers, simply pass "trust_remote_code=True":

```python
model_id = "arcee-ai/Trinity-Mini"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
    trust_remote_code=True
)
```

## VLLM

Supported in VLLM release 0.11.1

```
# pip
pip install "vllm>=0.11.1"
```

Serving the model with suggested settings:

```
vllm serve arcee-ai/Trinity-Mini \
  --dtype bfloat16 \
  --enable-auto-tool-choice \
  --reasoning-parser deepseek_r1 \
  --tool-call-parser hermes
```

## llama.cpp

Supported in llama.cpp release b7061

Download the latest [llama.cpp release](https://github.com/ggml-org/llama.cpp/releases)

```
llama-server -hf arcee-ai/Trinity-Mini-GGUF:q4_k_m \
  --temp 0.15 \
  --top-k 50 \
  --top-p 0.75
  --min-p 0.06
```

## LM Studio

Supported in latest LM Studio runtime

Update to latest available, then verify your runtime by:

1. Click "Power User" at the bottom left
2. Click the green "Developer" icon at the top left
3. Select "LM Runtimes" at the top
4. Refresh the list of runtimes and verify that the latest is installed

Then, go to Model Search and search for `arcee-ai/Trinity-Mini-GGUF`, download your prefered size, and load it up in the chat

## API

Trinity Mini is available today on openrouter:

https://openrouter.ai/arcee-ai/trinity-mini

```
curl -X POST "https://openrouter.ai/v1/chat/completions" \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "arcee-ai/trinity-mini",
    "messages": [
      {
        "role": "user",
        "content": "What are some fun things to do in New York?"
      }
    ]
  }'
```

## License

Trinity-Mini-W4A16 is released under the Apache-2.0 license.