mig-ai

Qwen3-Next-80B-A3B-Instruct

Deploy Dedicated

README

License: apache-2.0

Highlights

Qwen3-Next-80B-A3B is the first installment in the Qwen3-Next series and features the following key enchancements:

Hybrid Attention: Replaces standard attention with the combination of Gated DeltaNet and Gated Attention, enabling efficient context modeling for ultra-long context length.
High-Sparsity Mixture-of-Experts (MoE): Achieves an extreme low activation ratio in MoE layers, drastically reducing FLOPs per token while preserving model capacity.
Stability Optimizations: Includes techniques such as zero-centered and weight-decayed layernorm, and other stabilizing enhancements for robust pre-training and post-training.
Multi-Token Prediction (MTP): Boosts pretraining model performance and accelerates inference.

We are seeing strong performance in terms of both parameter efficiency and inference speed for Qwen3-Next-80B-A3B:

Qwen3-Next-80B-A3B-Base outperforms Qwen3-32B-Base on downstream tasks with 10% of the total training cost and with 10 times inference throughput for context over 32K tokens.
Qwen3-Next-80B-A3B-Instruct performs on par with Qwen3-235B-A22B-Instruct-2507 on certain benchmarks, while demonstrating significant advantages in handling ultra-long-context tasks up to 256K tokens.

Qwen3-Next-80B-A3B-Instruct Benchmark Comparison

For more details, please refer to our blog post Qwen3-Next.

Model Overview

[!Note] Qwen3-Next-80B-A3B-Instruct supports only instruct (non-thinking) mode and does not generate <think></think> blocks in its output.

Qwen3-Next-80B-A3B-Instruct has the following features:

Type: Causal Language Models
Training Stage: Pretraining (15T tokens) & Post-training
Number of Parameters: 80B in total and 3B activated
Number of Paramaters (Non-Embedding): 79B
Hidden Dimension: 2048
Number of Layers: 48
- Hybrid Layout: 12 * (3 * (Gated DeltaNet -> MoE) -> 1 * (Gated Attention -> MoE))
Gated Attention:
- Number of Attention Heads: 16 for Q and 2 for KV
- Head Dimension: 256
- Rotary Position Embedding Dimension: 64
Gated DeltaNet:
- Number of Linear Attention Heads: 32 for V and 16 for QK
- Head Dimension: 128

Performance

Table with columns: Qwen3-30B-A3B-Instruct-2507, Qwen3-32B Non-Thinking, Qwen3-235B-A22B-Instruct-2507, Qwen3-Next-80B-A3B-Instruct
	Qwen3-30B-A3B-Instruct-2507	Qwen3-32B Non-Thinking	Qwen3-235B-A22B-Instruct-2507	Qwen3-Next-80B-A3B-Instruct
Knowledge
MMLU-Pro	78.4	71.9	83.0	80.6
MMLU-Redux	89.3	85.7	93.1

*: For reproducibility, we report the win rates evaluated by GPT-4.1.

Quickstart

The code for Qwen3-Next has been merged into the main branch of Hugging Face transformers.

shell
pip install git+https://github.com/huggingface/transformers.git@main

With earlier versions, you will encounter the following error:

markdown
KeyError: 'qwen3_next'

The following contains a code snippet illustrating how to use the model generate content based on given inputs.

python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen/Qwen3-Next-80B-A3B-Instruct"

# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    dtype="auto",
    device_map="auto",
)

# prepare the model input
prompt = "Give me a short introduction to large language model."
messages = [
    {"role": "user", "content": prompt},
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=16384,
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 

content = tokenizer.decode(output_ids, skip_special_tokens=True)

print("content:", content)

[!Note] Multi-Token Prediction (MTP) is not generally available in Hugging Face Transformers.

[!Note] The efficiency or throughput improvement depends highly on the implementation. It is recommended to adopt a dedicated inference framework, e.g., SGLang and vLLM, for inference tasks.

[!Tip] Depending on the inference settings, you may observe better efficiency with flash-linear-attention and causal-conv1d. See the links for detailed instructions and requirements.

Deployment

For deployment, you can use the latest sglang or vllm to create an OpenAI-compatible API endpoint.

SGLang

SGLang is a fast serving framework for large language models and vision language models. SGLang could be used to launch a server with OpenAI-compatible API service.

sglang>=0.5.2 is required for Qwen3-Next, which can be installed using:

shell
pip install 'sglang[all]>=0.5.2'

See its documentation for more details.

The following command can be used to create an API endpoint at http://localhost:30000/v1 with maximum context length 256K tokens using tensor parallel on 4 GPUs.

shell
python -m sglang.launch_server --model-path Qwen/Qwen3-Next-80B-A3B-Instruct --port 30000 --tp-size 4 --context-length 262144 --mem-fraction-static 0.8

The following command is recommended for MTP with the rest settings the same as above:

shell
python -m sglang.launch_server --model-path Qwen/Qwen3-Next-80B-A3B-Instruct --port 30000 --tp-size 4 --context-length 262144 --mem-fraction-static 0.8 --speculative-algo NEXTN --speculative-num-steps 3 --speculative-eagle-topk 1 --speculative-num-draft-tokens 4

[!Note] The default context length is 256K. Consider reducing the context length to a smaller value, e.g., 32768, if the server fails to start.

Please also refer to SGLang's usage guide on Qwen3-Next.

vLLM

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. vLLM could be used to launch a server with OpenAI-compatible API service.

vllm>=0.10.2 is required for Qwen3-Next, which can be installed using:

shell
pip install 'vllm>=0.10.2'

See its documentation for more details.

The following command can be used to create an API endpoint at http://localhost:8000/v1 with maximum context length 256K tokens using tensor parallel on 4 GPUs.

shell
vllm serve Qwen/Qwen3-Next-80B-A3B-Instruct --port 8000 --tensor-parallel-size 4 --max-model-len 262144

The following command is recommended for MTP with the rest settings the same as above:

shell
vllm serve Qwen/Qwen3-Next-80B-A3B-Instruct --port 8000 --tensor-parallel-size 4 --max-model-len 262144 --speculative-config '{"method":"qwen3_next_mtp","num_speculative_tokens":2}'

[!Note] The default context length is 256K. Consider reducing the context length to a smaller value, e.g., 32768, if the server fails to start.

Please also refer to vLLM's usage guide on Qwen3-Next.

Agentic Use

Qwen3 excels in tool calling capabilities. We recommend using Qwen-Agent to make the best use of agentic ability of Qwen3. Qwen-Agent encapsulates tool-calling templates and tool-calling parsers internally, greatly reducing coding complexity.

To define the available tools, you can use the MCP configuration file, use the integrated tool of Qwen-Agent, or integrate other tools by yourself.

python
from qwen_agent.agents import Assistant

# Define LLM
llm_cfg = {
    'model': 'Qwen3-Next-80B-A3B-Instruct',

    # Use a custom endpoint compatible with OpenAI API:
    'model_server': 'http://localhost:8000/v1',  # api_base
    'api_key': 'EMPTY',
}

# Define Tools
tools = [
    {'mcpServers': {  # You can specify the MCP configuration file
            'time': {
                'command': 'uvx',
                'args': ['mcp-server-time', '--local-timezone=Asia/Shanghai']
            },
            "fetch": {
                "command": "uvx",
                "args": ["mcp-server-fetch"]
            }
        }
    },
  'code_interpreter',  # Built-in tools
]

# Define Agent
bot = Assistant(llm=llm_cfg, function_list=tools)

# Streaming generation
messages = [{'role': 'user', 'content': 'https://qwenlm.github.io/blog/ Introduce the latest developments of Qwen'}]
for responses in bot.run(messages=messages):
    pass
print(responses)

Processing Ultra-Long Texts

Qwen3-Next natively supports context lengths of up to 262,144 tokens. For conversations where the total length (including both input and output) significantly exceeds this limit, we recommend using RoPE scaling techniques to handle long texts effectively. We have validated the model's performance on context lengths of up to 1 million tokens using the YaRN method.

YaRN is currently supported by several inference frameworks, e.g., transformers, vllm and sglang. In general, there are two approaches to enabling YaRN for supported frameworks:

Modifying the model files: In the config.json file, add the rope_scaling fields:

json
{
    ...,
    "rope_scaling": {
        "rope_type": "yarn",
        "factor": 4.0,
        "original_max_position_embeddings": 262144
    }
}

Passing command line arguments:

For vllm, you can use

shell
VLLM_ALLOW_LONG_MAX_MODEL_LEN=1 vllm serve ... --rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":262144}' --max-model-len 1010000

For sglang, you can use

shell
SGLANG_ALLOW_OVERWRITE_LONGER_CONTEXT_LEN=1 python -m sglang.launch_server ... --json-model-override-args '{"rope_scaling":{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":262144}}' --context-length 1010000

[!NOTE] All the notable open-source frameworks implement static YaRN, which means the scaling factor remains constant regardless of input length, potentially impacting performance on shorter texts. We advise adding the rope_scaling configuration only when processing long contexts is required. It is also recommended to modify the factor as needed. For example, if the typical context length for your application is 524,288 tokens, it would be better to set factor as 2.0.

Long-Context Performance

We test the model on an 1M version of the RULER benchmark.

Table with columns: Model Name, Acc avg, 4k, 8k, 16k, 32k, 64k, 96k, 128k, 192k, 256k, 384k, 512k, 640k, 768k, 896k, 1000k
Model Name	Acc avg	4k	8k	16k	32k	64k	96k	128k	192k	256k	384k	512k	640k	768k	896k	1000k
Qwen3-30B-A3B-Instruct-2507	86.8	98.0

Qwen3-Next are evaluated with YaRN enabled. Qwen3-2507 models are evaluated with Dual Chunk Attention enabled.
Since the evaluation is time-consuming, we use 260 samples for each length (13 sub-tasks, 20 samples for each).

Best Practices

To achieve optimal performance, we recommend the following settings:

Sampling Parameters:
- We suggest using Temperature=0.7, TopP=0.8, TopK=20, and MinP=0.
- For supported frameworks, you can adjust the presence_penalty parameter between 0 and 2 to reduce endless repetitions. However, using a higher value may occasionally result in language mixing and a slight decrease in model performance.
Adequate Output Length: We recommend using an output length of 16,384 tokens for most queries, which is adequate for instruct models.
Standardize Output Format: We recommend using prompts to standardize model outputs when benchmarking.
- Math Problems: Include "Please reason step by step, and put your final answer within \boxed{}." in the prompt.
- : Add the following JSON structure to the prompt to standardize responses: "Please show your choice in the field with only the choice letter, e.g., ."

Citation

If you find our work helpful, feel free to give us a cite.

markdown
@misc{qwen3technicalreport,
      title={Qwen3 Technical Report}, 
      author={Qwen Team},
      year={2025},
      eprint={2505.09388},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2505.09388}, 
}

@article{qwen2.5-1m,
      title={Qwen2.5-1M Technical Report}, 
      author={An Yang and Bowen Yu and Chengyuan Li and Dayiheng Liu and Fei Huang and Haoyan Huang and Jiandong Jiang and Jianhong Tu and Jianwei Zhang and Jingren Zhou and Junyang Lin and Kai Dang and Kexin Yang and Le Yu and Mei Li and Minmin Sun and Qin Zhu and Rui Men and Tao He and Weijia Xu and Wenbiao Yin and Wenyuan Yu and Xiafei Qiu and Xingzhang Ren and Xinlong Yang and Yong Li and Zhiying Xu and Zipeng Zhang},
      journal={arXiv preprint arXiv:2501.15383},
      year={2025}
}

Available on FriendliAI

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Model Details

Model Provider

mig-ai

Model Tree

Base

this model

Input Modalities

Text

Output Modalities

Text

Supported Functionality

Dedicated Endpoints

Explore FriendliAI today

Get started Talk to an engineer

README

License: apache-2.0

Highlights

Qwen3-Next-80B-A3B is the first installment in the Qwen3-Next series and features the following key enchancements:

Hybrid Attention: Replaces standard attention with the combination of Gated DeltaNet and Gated Attention, enabling efficient context modeling for ultra-long context length.
High-Sparsity Mixture-of-Experts (MoE): Achieves an extreme low activation ratio in MoE layers, drastically reducing FLOPs per token while preserving model capacity.
Stability Optimizations: Includes techniques such as zero-centered and weight-decayed layernorm, and other stabilizing enhancements for robust pre-training and post-training.
Multi-Token Prediction (MTP): Boosts pretraining model performance and accelerates inference.

We are seeing strong performance in terms of both parameter efficiency and inference speed for Qwen3-Next-80B-A3B:

Qwen3-Next-80B-A3B-Base outperforms Qwen3-32B-Base on downstream tasks with 10% of the total training cost and with 10 times inference throughput for context over 32K tokens.
Qwen3-Next-80B-A3B-Instruct performs on par with Qwen3-235B-A22B-Instruct-2507 on certain benchmarks, while demonstrating significant advantages in handling ultra-long-context tasks up to 256K tokens.

Qwen3-Next-80B-A3B-Instruct Benchmark Comparison

For more details, please refer to our blog post Qwen3-Next.

Model Overview

[!Note] Qwen3-Next-80B-A3B-Instruct supports only instruct (non-thinking) mode and does not generate <think></think> blocks in its output.

Qwen3-Next-80B-A3B-Instruct has the following features:

Type: Causal Language Models
Training Stage: Pretraining (15T tokens) & Post-training
Number of Parameters: 80B in total and 3B activated
Number of Paramaters (Non-Embedding): 79B
Hidden Dimension: 2048
Number of Layers: 48
- Hybrid Layout: 12 * (3 * (Gated DeltaNet -> MoE) -> 1 * (Gated Attention -> MoE))
Gated Attention:
- Number of Attention Heads: 16 for Q and 2 for KV
- Head Dimension: 256
- Rotary Position Embedding Dimension: 64
Gated DeltaNet:
- Number of Linear Attention Heads: 32 for V and 16 for QK
- Head Dimension: 128

Performance

Table with columns: Qwen3-30B-A3B-Instruct-2507, Qwen3-32B Non-Thinking, Qwen3-235B-A22B-Instruct-2507, Qwen3-Next-80B-A3B-Instruct
	Qwen3-30B-A3B-Instruct-2507	Qwen3-32B Non-Thinking	Qwen3-235B-A22B-Instruct-2507	Qwen3-Next-80B-A3B-Instruct
Knowledge
MMLU-Pro	78.4	71.9	83.0	80.6
MMLU-Redux	89.3	85.7	93.1

*: For reproducibility, we report the win rates evaluated by GPT-4.1.

Quickstart

The code for Qwen3-Next has been merged into the main branch of Hugging Face transformers.

shell
pip install git+https://github.com/huggingface/transformers.git@main

With earlier versions, you will encounter the following error:

markdown
KeyError: 'qwen3_next'

The following contains a code snippet illustrating how to use the model generate content based on given inputs.

python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen/Qwen3-Next-80B-A3B-Instruct"

# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    dtype="auto",
    device_map="auto",
)

# prepare the model input
prompt = "Give me a short introduction to large language model."
messages = [
    {"role": "user", "content": prompt},
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=16384,
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 

content = tokenizer.decode(output_ids, skip_special_tokens=True)

print("content:", content)

[!Note] Multi-Token Prediction (MTP) is not generally available in Hugging Face Transformers.

[!Note] The efficiency or throughput improvement depends highly on the implementation. It is recommended to adopt a dedicated inference framework, e.g., SGLang and vLLM, for inference tasks.

[!Tip] Depending on the inference settings, you may observe better efficiency with flash-linear-attention and causal-conv1d. See the links for detailed instructions and requirements.

Deployment

For deployment, you can use the latest sglang or vllm to create an OpenAI-compatible API endpoint.

SGLang

SGLang is a fast serving framework for large language models and vision language models. SGLang could be used to launch a server with OpenAI-compatible API service.

sglang>=0.5.2 is required for Qwen3-Next, which can be installed using:

shell
pip install 'sglang[all]>=0.5.2'

See its documentation for more details.

The following command can be used to create an API endpoint at http://localhost:30000/v1 with maximum context length 256K tokens using tensor parallel on 4 GPUs.

shell
python -m sglang.launch_server --model-path Qwen/Qwen3-Next-80B-A3B-Instruct --port 30000 --tp-size 4 --context-length 262144 --mem-fraction-static 0.8

The following command is recommended for MTP with the rest settings the same as above:

shell
python -m sglang.launch_server --model-path Qwen/Qwen3-Next-80B-A3B-Instruct --port 30000 --tp-size 4 --context-length 262144 --mem-fraction-static 0.8 --speculative-algo NEXTN --speculative-num-steps 3 --speculative-eagle-topk 1 --speculative-num-draft-tokens 4

[!Note] The default context length is 256K. Consider reducing the context length to a smaller value, e.g., 32768, if the server fails to start.

Please also refer to SGLang's usage guide on Qwen3-Next.

vLLM

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. vLLM could be used to launch a server with OpenAI-compatible API service.

vllm>=0.10.2 is required for Qwen3-Next, which can be installed using:

shell
pip install 'vllm>=0.10.2'

See its documentation for more details.

The following command can be used to create an API endpoint at http://localhost:8000/v1 with maximum context length 256K tokens using tensor parallel on 4 GPUs.

shell
vllm serve Qwen/Qwen3-Next-80B-A3B-Instruct --port 8000 --tensor-parallel-size 4 --max-model-len 262144

The following command is recommended for MTP with the rest settings the same as above:

shell
vllm serve Qwen/Qwen3-Next-80B-A3B-Instruct --port 8000 --tensor-parallel-size 4 --max-model-len 262144 --speculative-config '{"method":"qwen3_next_mtp","num_speculative_tokens":2}'

[!Note] The default context length is 256K. Consider reducing the context length to a smaller value, e.g., 32768, if the server fails to start.

Please also refer to vLLM's usage guide on Qwen3-Next.

Agentic Use

To define the available tools, you can use the MCP configuration file, use the integrated tool of Qwen-Agent, or integrate other tools by yourself.

python
from qwen_agent.agents import Assistant

# Define LLM
llm_cfg = {
    'model': 'Qwen3-Next-80B-A3B-Instruct',

    # Use a custom endpoint compatible with OpenAI API:
    'model_server': 'http://localhost:8000/v1',  # api_base
    'api_key': 'EMPTY',
}

# Define Tools
tools = [
    {'mcpServers': {  # You can specify the MCP configuration file
            'time': {
                'command': 'uvx',
                'args': ['mcp-server-time', '--local-timezone=Asia/Shanghai']
            },
            "fetch": {
                "command": "uvx",
                "args": ["mcp-server-fetch"]
            }
        }
    },
  'code_interpreter',  # Built-in tools
]

# Define Agent
bot = Assistant(llm=llm_cfg, function_list=tools)

# Streaming generation
messages = [{'role': 'user', 'content': 'https://qwenlm.github.io/blog/ Introduce the latest developments of Qwen'}]
for responses in bot.run(messages=messages):
    pass
print(responses)

Processing Ultra-Long Texts

YaRN is currently supported by several inference frameworks, e.g., transformers, vllm and sglang. In general, there are two approaches to enabling YaRN for supported frameworks:

Modifying the model files: In the config.json file, add the rope_scaling fields:

json
{
    ...,
    "rope_scaling": {
        "rope_type": "yarn",
        "factor": 4.0,
        "original_max_position_embeddings": 262144
    }
}

Passing command line arguments:

For vllm, you can use

shell
VLLM_ALLOW_LONG_MAX_MODEL_LEN=1 vllm serve ... --rope-scaling '{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":262144}' --max-model-len 1010000

For sglang, you can use

shell
SGLANG_ALLOW_OVERWRITE_LONGER_CONTEXT_LEN=1 python -m sglang.launch_server ... --json-model-override-args '{"rope_scaling":{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":262144}}' --context-length 1010000

[!NOTE] All the notable open-source frameworks implement static YaRN, which means the scaling factor remains constant regardless of input length, potentially impacting performance on shorter texts. We advise adding the rope_scaling configuration only when processing long contexts is required. It is also recommended to modify the factor as needed. For example, if the typical context length for your application is 524,288 tokens, it would be better to set factor as 2.0.

Long-Context Performance

We test the model on an 1M version of the RULER benchmark.

Table with columns: Model Name, Acc avg, 4k, 8k, 16k, 32k, 64k, 96k, 128k, 192k, 256k, 384k, 512k, 640k, 768k, 896k, 1000k
Model Name	Acc avg	4k	8k	16k	32k	64k	96k	128k	192k	256k	384k	512k	640k	768k	896k	1000k
Qwen3-30B-A3B-Instruct-2507	86.8	98.0

Qwen3-Next are evaluated with YaRN enabled. Qwen3-2507 models are evaluated with Dual Chunk Attention enabled.
Since the evaluation is time-consuming, we use 260 samples for each length (13 sub-tasks, 20 samples for each).

Best Practices

To achieve optimal performance, we recommend the following settings:

Sampling Parameters:
- We suggest using Temperature=0.7, TopP=0.8, TopK=20, and MinP=0.
- For supported frameworks, you can adjust the presence_penalty parameter between 0 and 2 to reduce endless repetitions. However, using a higher value may occasionally result in language mixing and a slight decrease in model performance.
Adequate Output Length: We recommend using an output length of 16,384 tokens for most queries, which is adequate for instruct models.
Standardize Output Format: We recommend using prompts to standardize model outputs when benchmarking.
- Math Problems: Include "Please reason step by step, and put your final answer within \boxed{}." in the prompt.
- : Add the following JSON structure to the prompt to standardize responses: "Please show your choice in the field with only the choice letter, e.g., ."

Citation

If you find our work helpful, feel free to give us a cite.

markdown
@misc{qwen3technicalreport,
      title={Qwen3 Technical Report}, 
      author={Qwen Team},
      year={2025},
      eprint={2505.09388},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2505.09388}, 
}

@article{qwen2.5-1m,
      title={Qwen2.5-1M Technical Report}, 
      author={An Yang and Bowen Yu and Chengyuan Li and Dayiheng Liu and Fei Huang and Haoyan Huang and Jiandong Jiang and Jianhong Tu and Jianwei Zhang and Jingren Zhou and Junyang Lin and Kai Dang and Kexin Yang and Le Yu and Mei Li and Minmin Sun and Qin Zhu and Rui Men and Tao He and Weijia Xu and Wenbiao Yin and Wenyuan Yu and Xiafei Qiu and Xingzhang Ren and Xinlong Yang and Yong Li and Zhiying Xu and Zipeng Zhang},
      journal={arXiv preprint arXiv:2501.15383},
      year={2025}
}