---
title: AI21-Jamba-Reasoning-3B-6bit
canonical_url: "https://www.modelscope.cn/models/mlx-community/AI21-Jamba-Reasoning-3B-6bit"
md_url: "https://www.modelscope.cn/models/mlx-community/AI21-Jamba-Reasoning-3B-6bit.md"
repository: mlx-community/AI21-Jamba-Reasoning-3B-6bit
last_updated: 2025-10-14
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - jamba
architectures:
  - JambaForCausalLM
base_model:
  - ai21labs/AI21-Jamba-Reasoning-3B
base_model_relation: quantized
parameters: 681.8M
tensor_type:
  - U32
  - BF16
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
inference_backends:
  - "deploy_task text/emb"
  - "vllm 0.9.2"
downloads: 64
stars: 0
tags:
  - mlx
---

# AI21-Jamba-Reasoning-3B-6bit

> AI21-Jamba-Reasoning-3B-6bit - mlx-community 在 ModelScope 开源的模型。mlx-community/AI21-Jamba-Reasoning-3B-6bit

mlx-community/AI21-Jamba-Reasoning-3B-6bit 是 ModelScope 魔搭社区上的 681.8M 参数text-generation模型，采用 apache-2.0 许可，基于 ai21labs/AI21-Jamba-Reasoning-3B 构建，可用 deploy_task text/emb、vllm 0.9.2 部署。

- **Repository**: mlx-community/AI21-Jamba-Reasoning-3B-6bit
- **License**: apache-2.0
- **Tasks**: text-generation
- **Parameters**: 681.8M
- **Base model**: ai21labs/AI21-Jamba-Reasoning-3B
- **Inference backends**: deploy_task text/emb, vllm 0.9.2
- **Tags**: mlx
- **Downloads**: 64
- **Stars**: 0
- **Last updated**: 2025-10-14

Source: https://www.modelscope.cn/models/mlx-community/AI21-Jamba-Reasoning-3B-6bit

---

# mlx-community/AI21-Jamba-Reasoning-3B-6bit

This model [mlx-community/AI21-Jamba-Reasoning-3B-6bit](https://huggingface.co/mlx-community/AI21-Jamba-Reasoning-3B-6bit) was
converted to MLX format from [ai21labs/AI21-Jamba-Reasoning-3B](https://huggingface.co/ai21labs/AI21-Jamba-Reasoning-3B)
using mlx-lm version **0.28.2**.

## Use with mlx

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

```python
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/AI21-Jamba-Reasoning-3B-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)
```