Latest Content
Phi 4 Reasoning Plus icon

Phi 4 Reasoning Plus

NVIDIA
Phi-4-reasoning-plus is a dense transformer model optimized for advanced chain-of-thought reasoning across math, science, and coding tasks. It builds on Phi-4 with enhanced performance through supervised fine-tuning and outcome-based reinforcement learning for longer reasoning traces. The model features 14B parameters with 40 layers, 5,120 hidden size, and 40 attention heads with 10 key-value heads. It supports a 32K token context window and is trained on high-quality curated and synthetic data, delivering improved reasoning accuracy at the cost of higher latency.
TypeDense LLM
CapabilitiesText Generation, Instruction Following, Reasoning, Mathematical Reasoning+2 more
Group Release DateDecember 11, 2024
Links
Paper|Blog|HF Model Card
LicenseMIT

Inference Instructions

Deploy and run this model on NVIDIA B200 GPUs using the command below. Copy the command to get started with inference.

CONSOLE
docker run --gpus all 
 --shm-size 128g 
 -p 8000:8000 
 -v ~/.cache/huggingface:/root/.cache/huggingface 
 -e HF_TOKEN='YOUR_HF_TOKEN' 
 --ipc=host 
 lmsysorg/sglang:v0.5.9 
 python3 -m sglang.launch_server 
 --model-path microsoft/Phi-4-reasoning-plus 
 --host 0.0.0.0 
 --port 8000 
 --max-running-requests 1024 
 --max-prefill-tokens 65536 
 --tp 2 
 --enable-piecewise-cuda-graph 
 --mem-fraction-static 0.95 
 --trust-remote-code

Model Benchmarks

Each model was tested with a fixed input size and total token volume while increasing concurrency to measure serving performance under load.

ITL vs Concurrency

Time to First Token

Throughput Scaling

Total Tokens/sec vs Avg TTFT

Vultr Cloud GPU

NVIDIA HGX B200

Deploy NVIDIA B200 on Vultr Cloud GPU

Deploy NVIDIA B200 instancesContact Sales