Model ID
Qwen/Qwen3.5-2B
Capabilities
Qwen3.5 2B
Qwen3.5 2B is a compact multimodal causal language model built with a 2B parameter architecture. It uses a 24-layer transformer with 8 attention heads, 2 KV heads, and a 2,048 hidden size, paired with a 6,144 intermediate dimension. The model supports up to a 262K token context window and integrates a vision encoder with 24 layers and 1,024 hidden size. It combines Gated DeltaNet and attention layers, enabling efficient long-context reasoning, multimodal understanding, and strong multilingual performance across 200+ languages.
About this model
Qwen3.5 2B is a compact multimodal causal language model built with a 2B parameter architecture. It uses a 24-layer transformer with 8 attention heads, 2 KV heads, and a 2,048 hidden size, paired with a 6,144 intermediate dimension. The model supports up to a 262K token context window and integrates a vision encoder with 24 layers and 1,024 hidden size. It combines Gated DeltaNet and attention layers, enabling efficient long-context reasoning, multimodal understanding, and strong multilingual performance across 200+ languages.
Model specifications and capabilities are published by the model author and reproduced here from HF Model Card (Qwen/Qwen3.5-2B). Released under Apache 2.0. Further reading: Blog
Deploy Private Inference Endpoint
Choose your hardware and inference engine to get deployment commands and performance benchmarks tailored to your infrastructure
Hardware
MI325X
2,048 GB
B200
1,440 GB
Engine
vLLM
SGLang
Llama.cpp
Deployment code
Access Restricted
Enter your email to get access to this content
Performance under load
Tokens/s (total)
Benchmarks measured by Vultr on MI325X · vLLM. Throughput and latency vary with concurrency, input length and engine configuration, so treat these as a comparison baseline, not a service guarantee.
AMD Instinct MI325X
Unlock Next-Generation AI Power on Vultr Cloud GPU