As large language models (LLMs) grow larger, they get smarter, with open models from leading developers now featuring hundreds of billions of parameters. At the same time, today's leading models are also capable of reasoning, which means that they generate many intermediate reasoning tokens before delivering a final response to the user. The combination of these two trends-larger models that think using more tokens-drives the need for significantly higher compute performance. Delivering the highest performance on production workloads takes a state-of-the-art technology stack-spanning chips, systems, and software-and an expansive developer ecosystem that is constantly building on that stack.
MLPerf Inference v5.1 is the latest version of the MLPerf Inference industry standard benchmark. With benchmark rounds held twice per year, the benchmark features many tests of AI inference performance and is regularly updated with new models and scenarios. This round features:
DeepSeek-R1 - a popular 671-billion parameter mixture-of-experts (MoE) reasoning model, developed by DeepSeek. In the server scenario, the time-to-first-token (TTFT) threshold is 2 seconds with a 12.5 tokens/second/user (TPS/user) target. All TPS/user targets are 99th percentile, meaning that 99% of tokens meet or exceed that TPS/user speed.
Llama 3.1 405B - MLPerf Inference v5.1 adds a new interactive scenario for the largest of the Llama 3.1 series of models, providing a faster 12.5 TPS/user threshold with a shorter 4.5 second TTFT requirement compared to the existing server scenario.
Llama 3.1 8B - an 8-billion parameter member of the Llama 3.1 series of models with offline, server (2 second TTFT, 10 TPS/user), and interactive (0.5 second TTFT, 33 TPS/user) scenarios. This replaces the GPT-J benchmark used in prior rounds.
Whisper - a popular speech recognition model that recently saw nearly 5 million downloads in a month on HuggingFace. This replaces RNN-T, which was featured in prior editions of the MLPerf Inference benchmark suite.
This round, NVIDIA submitted the first results using the new Blackwell Ultra architecture, announced in March. It came just six months after Blackwell made its debut in the available category in MLPerf Inference v5.0, setting new inference performance records. Additionally, the NVIDIA platform set new performance records on all newly added benchmarks this round-DeepSeek-R1, Llama 3.1 405B, Llama 3.1 8B, and Whisper-and continues to hold per-GPU performance records on all other MLPerf inference benchmarks.
MLPerf Inference Per-Accelerator Records
Benchmark Offline Server Interactive
DeepSeek-R1 5,842 tokens/second/GPU 2,907 tokens/second/GPU **
Llama 3.1 405B 224 tokens/second/GPU 170 tokens/second/GPU 138 tokens/second/GPU
Llama 2 70B 99.9% 12,934 tokens/second/GPU 12,701 tokens/second/GPU 7,856 tokens/second/GPU
Llama 2 70B 99% 13,015 tokens/second/GPU 12,701 tokens/second/GPU 7,856 tokens/second/GPU
Llama 3.1 8B 18,370 tokens/second/GPU 16,099 tokens/second/GPU 15,284 tokens/second/GPU
Stable Diffusion XL 4.07 samples/second/GPU 3.59 queries/second/GPU **
Mixtral 8x7B 16,099 tokens/second/GPU 16,131 tokens/second/GPU **
DLRMv2 99% 87,228 samples/second/GPU 80,515 samples/second/GPU **
DLRMv2 99.9% 48,666 samples/second/GPU 46,259 queries/second/GPU **
Whisper 5,667 tokens/second/GPU ** **
R-GAT 81,404 samples/second/GPU ** **
Retinanet 1,875 samples/second/GPU 1,801 queries/second/GPU **
Table 1. Performance records per GPU based on submissions powered by the NVIDIA platform. MLPerf Inference v5.0 and v5.1, Closed Division. Results retrieved from www.mlcommons.org on September 9, 2025. NVIDIA platform results from the following entries: 5.0-0072, 5.1-0007, 5.1-0053, 5.1-0079, 5.1-0028, 5.1-0062, 5.1-0086, 5.1-0073, 5.1-0008, 5.1-0070,5.1-0046, 5.1-0009, 5.1-0060, 5.1-0072. 5.1-0071, 5.1-0069 Per chip performance derived by dividing total throughput by number of reported chips. Per-chip performance is not a primary metric of MLPerf Inference v5.0 or v5.1.The MLPerf name and logo are registered and unregistered trademarks of MLCommons Association in the United States and other countries. All rights reserved. Unauthorized use strictly prohibited. See www.mlcommons.org for more information.
NVIDIA also made extensive use of NVFP4 acceleration across all DeepSeek-R1 and Llama model submissions using the Blackwell and Blackwell Ultra architectures.
In this post, we take a closer look at these performance results and the full-stack technologies that enabled them.
Blackwell Ultra sets reasoning records in MLPerf debut This round, NVIDIA submitted results in the available category using the GB300 NVL72 rack-scale system, the first-ever MLPerf submissions using the Blackwell Ultra architecture. Blackwell Ultra builds upon the many advances in the NVIDIA Blackwell architecture, with several key enhancements:
1.5x higher peak NVFP4 AI compute
2x higher attention-layer compute
1.5x higher HBM3e capacity
Compared to the GB200 NVL72 submission, GB300 NVL72 delivered up to 45% higher performance per GPU, setting the standard on the new DeepSeek-R1 benchmark. And compared to unverified results collected on a Hopper-based system, Blackwell Ultra delivered about 5x higher throughput per GPU-translating into significantly higher AI factory throughput and much lower cost per token.
DeepSeek-R1 Performance
Architecture Offline Server
Hopper 1,253 tokens/second/GPU 556 tokens/second/GPU
Blackwell Ultra 5,842 tokens/second/GPU 2,907 tokens/second/GPU
Blackwell Ultra Advantage 4.7x 5.2x
Table 2. Per-GPU performance on DeepSeek-R1. MLPerf Inference v5.1, Closed. Blackwell Ultra results based on results in entry 5.1-0072. Hopper results not verified by MLCommons Association. Per-GPU performance is not a primary metric of MLPerf Inference v5.1 and is calcu
