Ornith-1.0-35B#

  • Context Length: 262144

  • Model Name: Ornith-1.0-35B

  • Languages: en, zh

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

  • Description: Ornith-1.0-35B is a 35B-total / 3B-activated Mixture-of-Experts multimodal model built on the Qwen3.5 MoE architecture (Qwen3_5MoeForConditionalGeneration). It combines hybrid linear/full attention, 256 routed experts (8 per token) plus a shared expert, multimodal RoPE, and multi-token prediction, with vision and video understanding via the Qwen3VL processor.

Specifications#

Model Spec 1 (pytorch, 35 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 35

  • Quantizations: none

  • Engines: vLLM, Transformers, SGLang

  • Model ID: deepreinforce-ai/Ornith-1.0-35B

  • 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 Ornith-1.0-35B --size-in-billions 35 --model-format pytorch --quantization ${quantization}