Model ID
deepseek-ai/DeepSeek-V4.1-Flash
Capabilities
DeepSeek V4.1 Flash
DeepSeek V4.1 Flash is a multimodal Mixture-of-Experts model designed for long-context reasoning, agentic workflows, and input-heavy workloads across text and images. It features 552B backbone parameters, using a 40-layer Causal Encoder-Decoder architecture with 5,120 hidden size, 64 attention heads, and 1 KV head. The model employs Compressed Sparse Attention 2, 384 routed experts with 6 activated per token, 1 shared expert, and 3 Multi-Token Prediction layers, alongside Engram conditional memory and DSpark speculative decoding. Supporting 1M-token context, it uses a 32-layer vision encoder for multimodal processing.
About this model
DeepSeek V4.1 Flash is a multimodal Mixture-of-Experts model designed for long-context reasoning, agentic workflows, and input-heavy workloads across text and images. It features 552B backbone parameters, using a 40-layer Causal Encoder-Decoder architecture with 5,120 hidden size, 64 attention heads, and 1 KV head. The model employs Compressed Sparse Attention 2, 384 routed experts with 6 activated per token, 1 shared expert, and 3 Multi-Token Prediction layers, alongside Engram conditional memory and DSpark speculative decoding. Supporting 1M-token context, it uses a 32-layer vision encoder for multimodal processing.
Model specifications and capabilities are published by the model author and reproduced here from Paper (deepseek-ai/DeepSeek-V4.1-Flash). Released under MIT. Further reading: Blog · HF Model Card
Deploy Private Inference Endpoint
Choose your hardware and inference engine to get deployment commands and performance benchmarks tailored to your infrastructure
Hardware
B200
1,440 GB
Engine
vLLM
SGLang
Deployment code
Access Restricted
Enter your email to get access to this content
Requires vllm/vllm-openai:deepseekv41-flash-0909 or any standard vLLM image v0.30.0+ to support this model configuration.
Performance under load
Tokens/s (total)
Benchmarks measured by Vultr on B200 · vLLM. Throughput and latency vary with concurrency, input length and engine configuration, so treat these as a comparison baseline, not a service guarantee.
NVIDIA HGX B200
Unlock Next-Generation AI Power on Vultr Cloud GPU