Latest Content
GPT OSS 120B icon

GPT OSS 120B

NVIDIA
Gpt-oss-120b is an open-weight mixture-of-experts (MoE) large language model built for high-performance reasoning, agentic workflows, and production-scale AI systems. It contains 117B total parameters with 5.1B active parameters per token, leveraging 128 experts (4 experts per token) across a 36-layer transformer with 64 attention heads. The model supports up to a 131K token context window with advanced RoPE scaling and sliding/full attention layers for efficient long-context processing.
TypeMoE LLM
CapabilitiesText Generation, Instruction Following, Reasoning, Mathematical Reasoning+5 more
ReleasedAugust 4, 2025
Links
Paper|Blog|HF Model Card
LicenseApache 2.0

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:v0.5.8-cu130 
 python3 -m sglang.launch_server 
 --model-path openai/gpt-oss-120b 
 --host 0.0.0.0 
 --port 8000 
 --max-prefill-tokens 65536 
 --max-running-requests 1024 
 --dtype auto 
 --tp 8 
 --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