---
title: INTELLECT-3-6bit
canonical_url: "https://www.modelscope.cn/models/mlx-community/INTELLECT-3-6bit"
md_url: "https://www.modelscope.cn/models/mlx-community/INTELLECT-3-6bit.md"
repository: mlx-community/INTELLECT-3-6bit
last_updated: 2025-11-27
license: mit
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - glm4_moe
architectures:
  - Glm4MoeForCausalLM
base_model:
  - PrimeIntellect/INTELLECT-3
base_model_relation: quantized
parameters: 20.7B
tensor_type:
  - U32
  - F32
  - BF16
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
inference_backends:
  - "deploy_task text"
  - "sglang 0.5.2"
downloads: 45
stars: 0
tags:
  - prime-rl
  - verifiers
  - prime-intellect
  - reinforcement-learning
  - reasoning
  - agentic
  - mixture-of-experts
  - mlx
---

# INTELLECT-3-6bit

> INTELLECT-3-6bit - mlx-community 在 ModelScope 开源的模型。mlx-community/INTELLECT-3-6bit

mlx-community/INTELLECT-3-6bit 是 ModelScope 魔搭社区上的 20.7B 参数text-generation模型，采用 mit 许可，基于 PrimeIntellect/INTELLECT-3 构建，可用 deploy_task text、sglang 0.5.2 部署。

- **Repository**: mlx-community/INTELLECT-3-6bit
- **License**: mit
- **Tasks**: text-generation
- **Parameters**: 20.7B
- **Base model**: PrimeIntellect/INTELLECT-3
- **Inference backends**: deploy_task text, sglang 0.5.2
- **Tags**: prime-rl, verifiers, prime-intellect, reinforcement-learning, reasoning, agentic, mixture-of-experts, mlx
- **Downloads**: 45
- **Stars**: 0
- **Last updated**: 2025-11-27

Source: https://www.modelscope.cn/models/mlx-community/INTELLECT-3-6bit

---

# mlx-community/INTELLECT-3-6bit

This model [mlx-community/INTELLECT-3-6bit](https://huggingface.co/mlx-community/INTELLECT-3-6bit) was
converted to MLX format from [PrimeIntellect/INTELLECT-3](https://huggingface.co/PrimeIntellect/INTELLECT-3)
using mlx-lm version **0.28.3**.

## Use with mlx

```bash
pip install mlx-lm
```

```python
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/INTELLECT-3-6bit")

prompt = "hello"

if tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)
```