huggit0000/Qwen3-0.6B-GGUF-FP32

Model Card

This is the FP32 GGUF version of Qwen3-0.6B. While the Qwen3 repository provides GGUF models, they are only available in quantized formats. In some cases, a full-precision (FP32) model is required. This version was therefore generated by converting the original Qwen3-0.6B safetensors model.

The converter in llama.cpp currently orders layers lexicographically, not by integer value. This version has all layers arranged sequentially in numerical order.

You can run this model on a single-file, pure C code directly without any dependencies. Visit this qwen3.c repo: https://github.com/gigit0000/qwen3.c

UPDATE: Now you can run on a single-file, pure CUDA C code without any dependencies - no cub, no cuBLAS. repo: https://github.com/gigit0000/qwen3.cu

Quick Start - Pure C version

// Compile and run
$ make run
$ ./run Qwen3-0.6B-FP32.gguf

// Or for faster inference with OpenMP
$ make runomp
$ OMP_NUM_THREADS=16 ./run Qwen3-0.6B-FP32.gguf  # the number of your cores

Multi-turn

$ OMP_NUM_THREADS=16 ./runb Qwen3-0.6B-FP32.gguf -m 0 -k 1
Multi-turn = on, thinKing = off, Temperature = 0.60, top-P = 0.95
Strike Enter when you wanna exit this chat
Enter system prompt (or Enter to skip): Tell me in one sentence
Q: Where is the best spot in Paris?
A: The best spot in Paris is the Eiffel Tower.
Q: What about the second-best spot?
A: The second-best spot in Paris is the Louvre Museum.

Reasoning

$ OMP_NUM_THREADS=16 ./run Qwen3-0.6B-FP32.gguf -k 0
Multi-turn = off, thinKing = on, Temperature = 0.60, top-P = 0.95
Strike Enter when you wanna exit this chat
Enter system prompt (or Enter to skip): 
Q: Why do stars shine? Give me a quick answer!
A: <think>
Okay, the user is asking why stars shine. Let me start by recalling what I know about stars. Stars are luminous objects that emit light. So, the main reason they shine is because they produce light through nuclear fusion.

Wait, but I should make sure. Stars form from clouds of gas and dust in space. When these clouds cool, they start fusing hydrogen into helium, which releases energy. This energy is what we see as light. So the process is nuclear fusion of hydrogen into helium, which gives off energy.

I should also mention that the energy from stars is what we perceive as light. Maybe add that this light travels through space and we see it on Earth. But the question is why they shine, so the answer should focus on the energy production.

I need to keep it simple and concise. The user probably wants a quick answer, so no need for too much detail. Let me check if there's any other reason, but I think that's the main one. Alright, I think that's it.
</think>

Stars shine because they produce light through nuclear fusion of hydrogen into helium in their cores. This energy is then released as visible light, giving them their luminous glow.

Quick Start - Pure CUDA C

$ make runcu
$ ./runcu Qwen3-0.6B-FP32.gguf

All other options work the same in the C version.

You can monitor TPS with -r option.

$ ./runcu Qwen3-0.6B-FP32.gguf -r 1 
Multi-turn = off, thinKing = off, tps(R) = on, Temperature = 0.60, top-P = 0.95
Press Enter to exit the chat
Enter system prompt (or Enter to skip): You name is Tom.
Q: What is your name?
A: My name is Tom.
tok/s: 34.482759

Enjoy it!

=====================================================

Model Card Copied from Original Repo

Qwen3 Highlights Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Built upon extensive training, Qwen3 delivers groundbreaking advancements in reasoning, instruction-following, agent capabilities, and multilingual support, with the following key features:

Uniquely support of seamless switching between thinking mode (for complex logical reasoning, math, and coding) and non-thinking mode (for efficient, general-purpose dialogue) within single model, ensuring optimal performance across various scenarios. Significantly enhancement in its reasoning capabilities, surpassing previous QwQ (in thinking mode) and Qwen2.5 instruct models (in non-thinking mode) on mathematics, code generation, and commonsense logical reasoning. Superior human preference alignment, excelling in creative writing, role-playing, multi-turn dialogues, and instruction following, to deliver a more natural, engaging, and immersive conversational experience. Expertise in agent capabilities, enabling precise integration with external tools in both thinking and unthinking modes and achieving leading performance among open-source models in complex agent-based tasks. Support of 100+ languages and dialects with strong capabilities for multilingual instruction following and translation. Model Overview Qwen3-0.6B has the following features:

Type: Causal Language Models Training Stage: Pretraining & Post-training Number of Parameters: 0.6B Number of Paramaters (Non-Embedding): 0.44B Number of Layers: 28 Number of Attention Heads (GQA): 16 for Q and 8 for KV Context Length: 32,768 For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our blog, GitHub, and Documentation.

If you encounter significant endless repetitions, please refer to the Best Practices section for optimal sampling parameters, and set the presence_penalty to 1.5.

Quickstart The code of Qwen3 has been in the latest Hugging Face transformers and we advise you to use the latest version of transformers.

With transformers<4.51.0, you will encounter the following error:

KeyError: 'qwen3'

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

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen/Qwen3-0.6B"

load the tokenizer and the model

tokenizer = AutoTokenizer.from_pretrained(model_name) model = AutoModelForCausalLM.from_pretrained( model_name, torch_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, enable_thinking=True # Switches between thinking and non-thinking modes. Default is True. ) model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

conduct text completion

generated_ids = model.generate( **model_inputs, max_new_tokens=32768 ) output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist()

parsing thinking content

try: # rindex finding 151668 () index = len(output_ids) - output_ids[::-1].index(151668) except ValueError: index = 0

thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n") content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")

print("thinking content:", thinking_content) print("content:", content)

For deployment, you can use sglang>=0.4.6.post1 or vllm>=0.8.5 or to create an OpenAI-compatible API endpoint:

SGLang: python -m sglang.launch_server --model-path Qwen/Qwen3-0.6B --reasoning-parser qwen3

vLLM: vllm serve Qwen/Qwen3-0.6B --enable-reasoning --reasoning-parser deepseek_r1

For local use, applications such as Ollama, LMStudio, MLX-LM, llama.cpp, and KTransformers have also supported Qwen3.

Switching Between Thinking and Non-Thinking Mode The enable_thinking switch is also available in APIs created by SGLang and vLLM. Please refer to our documentation for SGLang and vLLM users.

enable_thinking=True By default, Qwen3 has thinking capabilities enabled, similar to QwQ-32B. This means the model will use its reasoning abilities to enhance the quality of generated responses. For example, when explicitly setting enable_thinking=True or leaving it as the default value in tokenizer.apply_chat_template, the model will engage its thinking mode.

text = tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True, enable_thinking=True # True is the default value for enable_thinking )

In this mode, the model will generate think content wrapped in a ... block, followed by the final response.

For thinking mode, use Temperature=0.6, TopP=0.95, TopK=20, and MinP=0 (the default setting in generation_config.json). DO NOT use greedy decoding, as it can lead to performance degradation and endless repetitions. For more detailed guidance, please refer to the Best Practices section.

enable_thinking=False We provide a hard switch to strictly disable the model's thinking behavior, aligning its functionality with the previous Qwen2.5-Instruct models. This mode is particularly useful in scenarios where disabling thinking is essential for enhancing efficiency.

text = tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True, enable_thinking=False # Setting enable_thinking=False disables thinking mode )

In this mode, the model will not generate any think content and will not include a ... block.

For non-thinking mode, we suggest using Temperature=0.7, TopP=0.8, TopK=20, and MinP=0. For more detailed guidance, please refer to the Best Practices section.

Advanced Usage: Switching Between Thinking and Non-Thinking Modes via User Input We provide a soft switch mechanism that allows users to dynamically control the model's behavior when enable_thinking=True. Specifically, you can add /think and /no_think to user prompts or system messages to switch the model's thinking mode from turn to turn. The model will follow the most recent instruction in multi-turn conversations.

Here is an example of a multi-turn conversation:

from transformers import AutoModelForCausalLM, AutoTokenizer

class QwenChatbot: def init(self, model_name="Qwen/Qwen3-0.6B"): self.tokenizer = AutoTokenizer.from_pretrained(model_name) self.model = AutoModelForCausalLM.from_pretrained(model_name) self.history = []

def generate_response(self, user_input):
    messages = self.history + [{"role": "user", "content": user_input}]

    text = self.tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True
    )

    inputs = self.tokenizer(text, return_tensors="pt")
    response_ids = self.model.generate(**inputs, max_new_tokens=32768)[0][len(inputs.input_ids[0]):].tolist()
    response = self.tokenizer.decode(response_ids, skip_special_tokens=True)

    # Update history
    self.history.append({"role": "user", "content": user_input})
    self.history.append({"role": "assistant", "content": response})

    return response

Example Usage

if name == "main": chatbot = QwenChatbot()

# First input (without /think or /no_think tags, thinking mode is enabled by default)
user_input_1 = "How many r's in strawberries?"
print(f"User: {user_input_1}")
response_1 = chatbot.generate_response(user_input_1)
print(f"Bot: {response_1}")
print("----------------------")

# Second input with /no_think
user_input_2 = "Then, how many r's in blueberries? /no_think"
print(f"User: {user_input_2}")
response_2 = chatbot.generate_response(user_input_2)
print(f"Bot: {response_2}") 
print("----------------------")

# Third input with /think
user_input_3 = "Really? /think"
print(f"User: {user_input_3}")
response_3 = chatbot.generate_response(user_input_3)
print(f"Bot: {response_3}")

For API compatibility, when enable_thinking=True, regardless of whether the user uses /think or /no_think, the model will always output a block wrapped in .... However, the content inside this block may be empty if thinking is disabled. When enable_thinking=False, the soft switches are not valid. Regardless of any /think or /no_think tags input by the user, the model will not generate think content and will not include a ... block.

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.

from qwen_agent.agents import Assistant

Define LLM

llm_cfg = { 'model': 'Qwen3-0.6B',

# Use the endpoint provided by Alibaba Model Studio:
# 'model_type': 'qwen_dashscope',
# 'api_key': os.getenv('DASHSCOPE_API_KEY'),

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

# Other parameters:
# 'generate_cfg': {
#         # Add: When the response content is `<think>this is the thought</think>this is the answer;
#         # Do not add: When the response has been separated by reasoning_content and content.
#         'thought_in_content': True,
#     },

}

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)

Best Practices To achieve optimal performance, we recommend the following settings:

Sampling Parameters:

For thinking mode (enable_thinking=True), use Temperature=0.6, TopP=0.95, TopK=20, and MinP=0. DO NOT use greedy decoding, as it can lead to performance degradation and endless repetitions. For non-thinking mode (enable_thinking=False), 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 32,768 tokens for most queries. For benchmarking on highly complex problems, such as those found in math and programming competitions, we suggest setting the max output length to 38,912 tokens. This provides the model with sufficient space to generate detailed and comprehensive responses, thereby enhancing its overall performance.

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. Multiple-Choice Questions: Add the following JSON structure to the prompt to standardize responses: "Please show your choice in the answer field with only the choice letter, e.g., "answer": "C"." No Thinking Content in History: In multi-turn conversations, the historical model output should only include the final output part and does not need to include the thinking content. It is implemented in the provided chat template in Jinja2. However, for frameworks that do not directly use the Jinja2 chat template, it is up to the developers to ensure that the best practice is followed.

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

@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}, }