---
title: Qwen3-14B-Instruct-Exp
canonical_url: "https://www.modelscope.cn/models/OpenPipe/Qwen3-14B-Instruct-Exp"
md_url: "https://www.modelscope.cn/models/OpenPipe/Qwen3-14B-Instruct-Exp.md"
repository: OpenPipe/Qwen3-14B-Instruct-Exp
last_updated: 2025-10-10
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen3
architectures:
  - Qwen3ForCausalLM
base_model:
  - Qwen/Qwen3-14B-Base
base_model_relation: finetune
parameters: 14.8B
tensor_type:
  - BF16
library_name:
  - safetensors
  - pytorch
frameworks:
  - pytorch
inference_backends:
  - "deploy_task text/emb"
  - "lmdeploy_turbomind 0.9.1"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 10
stars: 0
---

# Qwen3-14B-Instruct-Exp

> Qwen3-14B-Instruct-Exp - OpenPipe 在 ModelScope 开源的模型。Qwen3-14B-Instruct Highlights

OpenPipe/Qwen3-14B-Instruct-Exp 是 ModelScope 魔搭社区上的 14.8B 参数text-generation模型，采用 apache-2.0 许可，基于 Qwen/Qwen3-14B-Base 构建，可用 deploy_task text/emb、lmdeploy_turbomind 0.9.1、sglang 0.5.2 部署。

- **Repository**: OpenPipe/Qwen3-14B-Instruct-Exp
- **License**: apache-2.0
- **Tasks**: text-generation
- **Parameters**: 14.8B
- **Base model**: Qwen/Qwen3-14B-Base
- **Inference backends**: deploy_task text/emb, lmdeploy_turbomind 0.9.1, sglang 0.5.2, vllm 0.9.2
- **Downloads**: 10
- **Stars**: 0
- **Last updated**: 2025-10-10

Source: https://www.modelscope.cn/models/OpenPipe/Qwen3-14B-Instruct-Exp

---

# Qwen3-14B
<a href="https://chat.qwen.ai/" target="_blank" style="margin: 2px;">
    <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/>
</a>

## Qwen3-14B-Instruct Highlights

OpenPipe/Qwen3-14B-Instruct is a finetune friendly instruct variant of Qwen3-14B. Qwen3 release does not include a 14B Instruct (non-thinking) model, this fork introduces an updated chat template that makes Qwen3-14B non-thinking by default and be highly compatible with OpenPipe and other finetuning frameworks. The default Qwen3 chat template does not render `<think></think>` tags on the previous assistant message, which can lead to inconsistencies between training and generation. This version resolves that issue by adding `<think></think>` tags to all assistant prompts and generation templates to ensure message format consistency during both training and inference. The model retains the strong general capabilities of Qwen3-14B while providing a more finetuning friendly chat template.

## Model Overview

**Qwen3-14B** has the following features:
- Type: Causal Language Models
- Training Stage: Pretraining & Post-training
- Number of Parameters: 14.8B
- Number of Paramaters (Non-Embedding): 13.2B
- Number of Layers: 40
- Number of Attention Heads (GQA): 40 for Q and 8 for KV
- Context Length: 32,768 natively and [131,072 tokens with YaRN](#processing-long-texts). 

For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our [blog](https://qwenlm.github.io/blog/qwen3/), [GitHub](https://github.com/QwenLM/Qwen3), and [Documentation](https://qwen.readthedocs.io/en/latest/).