Top
Author
Published
15 Aug 2025Form Number
LP2277PDF size
5 pages, 288 KBSubscribed to LP2277.
Thank you for your feedback.
Abstract
The release of MLPerf Training 5.0 benchmarks by MLCommons marks a significant milestone in the evaluation of machine learning performance. These comprehensive benchmarks provide insights into AI system capabilities, focusing on both training and inference across diverse hardware configurations. In this article, we will explore our contributions to MLPerf Training 5.0 with our latest Lenovo ThinkSystem SR680a V3 and SR780a V3 systems, powered by the latest NVIDIA H200 and B200 GPUs and we’ll discuss the industry applications of these impressive results.
Introduction
MLPerf is the industry’s gold standard for measuring AI performance, developed by a collaboration of leading tech companies and academic institutions. It benchmarks real-world AI workloads like large language models, computer vision, and more. With the release of MLPerf Training 5.0, Lenovo has once again proven its strength in AI infrastructure delivering standout results.
Lenovo continues to prove its leadership in AI infrastructure with a strong showing in the latest MLPerf Training 5.0 benchmarks. Lenovo ThinkSystem SR680a V3 and SR780a V3 systems delivered impressive performance across a range of AI training workloads—including large language models, generative AI, and computer vision. This is the first Lenovo submission with a multi-node cluster: 4 nodes SR680a V3 with 8x H200-SXM-141GB in each.
Highlights from MLPerf Training 5.0
Lenovo proudly participated in the MLPerf Training 5.0 benchmark suite, achieving outstanding results with the state-of-the-art Lenovo ThinkSystem SR680a V3 and SR780a V3 systems, with the latest NVIDIA H200 and B200 GPUs.
ThinkSystem SR680a V3 with 32x H200-SXM-141GB (4 Nodes, 8 GPUs in each)
Here are the key outcomes with the ThinkSystem SR680a V3:
- LLMs - LaMA2-70B-LoRA: Training Time: 6.94 minutes
- Computer Vision - RetinaNet: Training Time: 9.64 minutes
- Generative AI - Stable Diffusion (Training): Training Time: 79 minutes
The Lenovo ThinkSystem SR680a V3 with 32x H200-SXM-141GB (4 Nodes, 8 GPUs in each) is an NVIDIA Certified system built for large-scale AI training, the SR680a V3 showed its muscle across some of the most demanding benchmarks earning 1st place in LLMs.
These results confirm the SR680a V3’s capability to handle advanced AI workloads at scale—ideal for organizations training cutting-edge models that require massive compute.
Lenovo ThinkSystem SR780a V3 with 8x B200-SXM-180GB (Single Node)
Performance results for the powerful Lenovo ThinkSystem SR780a V3 have been officially published and can be viewed on the MLCommons website, showcasing its capabilities in the industry-standard MLPerf Training benchmarks:
https://mlcommons.org/benchmarks/training/
Why Lenovo?
Lenovo’s strong benchmark performance is no accident—it’s the result of thoughtful design and a clear focus on what AI users need most.
Here are a few key reasons why Lenovo systems continue to stand out:
- Scalability That Fits Your Needs
Whether you’re training massive LLMs or building models on a single node, Lenovo systems offer the flexibility to grow with your AI projects. - Versatility Across Workloads
From computer vision and generative AI to fine-tuning large models with LoRA, Lenovo delivers consistent, dependable performance. - Powered by the Best
Equipped with the latest NVIDIA H200 and B200 GPUs, ThinkSystem platforms are built to tackle today’s toughest AI challenges with speed and efficiency.
A Future-Focused Approach to AI
The MLPerf Training 5.0 benchmarks reinforces the rapid evolution of AI hardware and software, highlighting the potential for significant improvements in performance and efficiency. Our contributions to these benchmarks reflect our commitment to innovation and pushing the boundaries achieving business outcomes with AI.
As we continue to innovate, the insights gained from MLPerf Training 5.0 will guide us in developing cutting-edge AI solutions that address the most pressing challenges across industries. The transformative power of AI is evolving fast and is becoming increasingly evident. Whether its supporting cutting-edge research, helping businesses deploy AI at scale, or building tools for the next generation of generative models, Lenovo is all-in on powering real progress.
We are excited to be at the forefront of this revolution.
Conclusion
Lenovo’s latest results in the MLPerf Training 5.0 benchmarks highlight the strength and versatility of its AI systems. With the SR680a V3 leading in large-scale training and the SR780a V3 offering a compact and powerful solution, Lenovo is ready to support organizations at every stage of their AI journey. Backed by powerful hardware, smart design, and proven performance, Lenovo continues to be a reliable partner for teams building and scaling the future of AI.
Professionals in the field are encouraged to consider these results in their future hardware procurement and system design strategies.
Author
Traci Parker is the Worldwide Solutions Marketing Manager for Enterprise IT and AI at Lenovo. She specializes in hybrid cloud, infrastructure modernization and AI solutions. She has more than 15 years of experience as a Marketing Manager and Product Marketing Manager across high-tech, fin-tech and healthcare industries.
Trademarks
Lenovo and the Lenovo logo are trademarks or registered trademarks of Lenovo in the United States, other countries, or both. A current list of Lenovo trademarks is available on the Web at https://www.lenovo.com/us/en/legal/copytrade/.
The following terms are trademarks of Lenovo in the United States, other countries, or both:
Lenovo®
ThinkSystem®
The following terms are trademarks of other companies:
Approach® is a trademark of IBM in the United States, other countries, or both.
Other company, product, or service names may be trademarks or service marks of others.
Configure and Buy
Please select a locale
Full Change History
Course Detail
Employees Only Content
The content in this document with a is only visible to employees who are logged in. Logon using your Lenovo ITcode and password via Lenovo single-signon (SSO).
The author of the document has determined that this content is classified as Lenovo Internal and should not be normally be made available to people who are not employees or contractors. This includes partners, customers, and competitors. The reasons may vary and you should reach out to the authors of the document for clarification, if needed. Be cautious about sharing this content with others as it may contain sensitive information.
Any visitor to the Lenovo Press web site who is not logged on will not be able to see this employee-only content. This content is excluded from search engine indexes and will not appear in any search results.
For all users, including logged-in employees, this employee-only content does not appear in the PDF version of this document.
This functionality is cookie based. The web site will normally remember your login state between browser sessions, however, if you clear cookies at the end of a session or work in an Incognito/Private browser window, then you will need to log in each time.
If you have any questions about this feature of the Lenovo Press web, please email David Watts at dwatts@lenovo.com.
