Qwen3-Thinking#

  • Context Length: 262144

  • Model Name: Qwen3-Thinking

  • Languages: en, zh

  • Abilities: chat, reasoning, tools

  • Description: we have continued to scale the thinking capability of Qwen3-235B-A22B, improving both the quality and depth of reasoning

Specifications#

Model Spec 1 (pytorch, 235 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 235

  • Quantizations: none

  • Engines: vLLM, Transformers

  • Model ID: Qwen/Qwen3-235B-A22B-Thinking-2507

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 235 --model-format pytorch --quantization ${quantization}

Model Spec 2 (pytorch, 30 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 30

  • Quantizations: none

  • Engines: vLLM, Transformers

  • Model ID: Qwen/Qwen3-30B-A3B-Thinking-2507

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 30 --model-format pytorch --quantization ${quantization}

Model Spec 3 (pytorch, 4 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 4

  • Quantizations: none

  • Engines: vLLM, Transformers

  • Model ID: Qwen/Qwen3-4B-Thinking-2507

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 4 --model-format pytorch --quantization ${quantization}

Model Spec 4 (fp8, 235 Billion)#

  • Model Format: fp8

  • Model Size (in billions): 235

  • Quantizations: fp8

  • Engines: vLLM

  • Model ID: Qwen/Qwen3-235B-A22B-Thinking-2507-FP8

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 235 --model-format fp8 --quantization ${quantization}

Model Spec 5 (fp8, 30 Billion)#

  • Model Format: fp8

  • Model Size (in billions): 30

  • Quantizations: fp8

  • Engines: vLLM

  • Model ID: Qwen/Qwen3-30B-A3B-Thinking-2507-FP8

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 30 --model-format fp8 --quantization ${quantization}

Model Spec 6 (fp8, 4 Billion)#

  • Model Format: fp8

  • Model Size (in billions): 4

  • Quantizations: none

  • Engines: vLLM

  • Model ID: Qwen/Qwen3-4B-Thinking-2507-FP8

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 4 --model-format fp8 --quantization ${quantization}

Model Spec 7 (gptq, 235 Billion)#

  • Model Format: gptq

  • Model Size (in billions): 235

  • Quantizations: Int4-Int8Mix

  • Engines: vLLM, Transformers

  • Model ID: QuantTrio/Qwen3-235B-A22B-Thinking-2507-GPTQ-Int4-Int8Mix

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 235 --model-format gptq --quantization ${quantization}

Model Spec 8 (gptq, 30 Billion)#

  • Model Format: gptq

  • Model Size (in billions): 30

  • Quantizations: Int8

  • Engines: vLLM, Transformers

  • Model ID: QuantTrio/Qwen3-30B-A3B-Thinking-2507-GPTQ-Int8

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 30 --model-format gptq --quantization ${quantization}

Model Spec 9 (gptq, 4 Billion)#

  • Model Format: gptq

  • Model Size (in billions): 4

  • Quantizations: Int4, Int8

  • Engines: vLLM, Transformers

  • Model ID: JunHowie/Qwen3-4B-Thinking-2507-GPTQ-{quantization}

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 4 --model-format gptq --quantization ${quantization}

Model Spec 10 (awq, 235 Billion)#

  • Model Format: awq

  • Model Size (in billions): 235

  • Quantizations: Int4

  • Engines: vLLM, Transformers

  • Model ID: QuantTrio/Qwen3-235B-A22B-Thinking-2507-AWQ

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 235 --model-format awq --quantization ${quantization}

Model Spec 11 (awq, 30 Billion)#

  • Model Format: awq

  • Model Size (in billions): 30

  • Quantizations: Int4

  • Engines: vLLM, Transformers

  • Model ID: QuantTrio/Qwen3-30B-A3B-Thinking-2507-AWQ

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 30 --model-format awq --quantization ${quantization}

Model Spec 12 (awq, 4 Billion)#

  • Model Format: awq

  • Model Size (in billions): 4

  • Quantizations: Int4

  • Engines: vLLM, Transformers

  • Model ID: Eslzzyl/Qwen3-4B-Thinking-2507-AWQ

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 4 --model-format awq --quantization ${quantization}

Model Spec 13 (mlx, 235 Billion)#

  • Model Format: mlx

  • Model Size (in billions): 235

  • Quantizations: 3bit, 4bit, 5bit, 8bit

  • Engines: MLX

  • Model ID: mlx-community/Qwen3-235B-A22B-Thinking-2507-{quantization}

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 235 --model-format mlx --quantization ${quantization}

Model Spec 14 (mlx, 30 Billion)#

  • Model Format: mlx

  • Model Size (in billions): 30

  • Quantizations: 4bit, bf16

  • Engines: MLX

  • Model ID: mlx-community/Qwen3-30B-A3B-Thinking-2507-{quantization}

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 30 --model-format mlx --quantization ${quantization}

Model Spec 15 (mlx, 4 Billion)#

  • Model Format: mlx

  • Model Size (in billions): 4

  • Quantizations: 4bit, 5bit, 6bit, 8bit

  • Engines: MLX

  • Model ID: mlx-community/Qwen3-4B-Thinking-2507-{quantization}

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 4 --model-format mlx --quantization ${quantization}

Model Spec 16 (ggufv2, 235 Billion)#

  • Model Format: ggufv2

  • Model Size (in billions): 235

  • Quantizations: BF16, IQ4_XS, Q2_K, Q2_K_L, Q3_K_M, Q3_K_S, Q4_0, Q4_1, Q4_K_M, Q4_K_S, Q5_K_M, Q5_K_S, Q6_K, Q8_0, UD-Q2_K_XL, UD-Q3_K_XL, UD-Q4_K_XL, UD-Q5_K_XL, UD-Q6_K_XL, UD-Q8_K_XL

  • Engines: vLLM, llama.cpp

  • Model ID: unsloth/Qwen3-235B-A22B-Thinking-2507-GGUF

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 235 --model-format ggufv2 --quantization ${quantization}

Model Spec 17 (ggufv2, 30 Billion)#

  • Model Format: ggufv2

  • Model Size (in billions): 30

  • Quantizations: BF16, IQ4_NL, IQ4_XS, Q2_K, Q2_K_L, Q3_K_M, Q3_K_S, Q4_0, Q4_1, Q4_K_M, Q4_K_S, Q5_K_M, Q5_K_S, Q6_K, Q8_0, UD-IQ1_M, UD-IQ1_S, UD-IQ2_M, UD-IQ2_XXS, UD-IQ3_XXS, UD-Q2_K_XL, UD-Q3_K_XL, UD-Q4_K_XL, UD-Q5_K_XL, UD-Q6_K_XL, UD-Q8_K_XL, UD-TQ1_0

  • Engines: vLLM, llama.cpp

  • Model ID: unsloth/Qwen3-30B-A3B-Thinking-2507-GGUF

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 30 --model-format ggufv2 --quantization ${quantization}

Model Spec 18 (ggufv2, 4 Billion)#

  • Model Format: ggufv2

  • Model Size (in billions): 4

  • Quantizations: BF16, IQ4_NL, IQ4_XS, Q2_K, Q2_K_L, Q3_K_M, Q3_K_S, Q4_0, Q4_1, Q4_K_M, Q4_K_S, Q5_K_M, Q5_K_S, Q6_K, Q8_0, UD-IQ1_M, UD-IQ1_S, UD-IQ2_M, UD-IQ2_XXS, UD-IQ3_XXS, UD-Q2_K_XL, UD-Q3_K_XL, UD-Q4_K_XL, UD-Q5_K_XL, UD-Q6_K_XL, UD-Q8_K_XL

  • Engines: vLLM, llama.cpp

  • Model ID: unsloth/Qwen3-4B-Thinking-2507-GGUF

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name Qwen3-Thinking --size-in-billions 4 --model-format ggufv2 --quantization ${quantization}