Nex-N2#

  • Context Length: 262144

  • Model Name: Nex-N2

  • Languages: en, zh

  • Abilities: chat, vision, tools, reasoning, hybrid

  • Description: Nex-N2 is a series of multimodal large language models developed by nex-agi, built on the Qwen3.5 MoE architecture. It supports text, image, and video understanding, with advanced reasoning and tool-use capabilities.

Specifications#

Model Spec 1 (pytorch, 35 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 35

  • Quantizations: none

  • Engines: vLLM, Transformers, SGLang

  • Model ID: nex-agi/Nex-N2-mini

  • 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 Nex-N2 --size-in-billions 35 --model-format pytorch --quantization ${quantization}

Model Spec 2 (pytorch, 397 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 397

  • Quantizations: none

  • Engines: vLLM, Transformers, SGLang

  • Model ID: nex-agi/Nex-N2-Pro

  • 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 Nex-N2 --size-in-billions 397 --model-format pytorch --quantization ${quantization}

Model Spec 3 (fp8, 397 Billion)#

  • Model Format: fp8

  • Model Size (in billions): 397

  • Quantizations: FP8

  • Engines: vLLM, Transformers, SGLang

  • Model ID: nex-agi/Nex-N2-Pro-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 Nex-N2 --size-in-billions 397 --model-format fp8 --quantization ${quantization}