Congratulations to MLPerf Inference v6.1 submitters - results are live!!!
David Kanter
Hi Everyone,
Good morning! It is my delight to announce the release of MLPerf Inference v6.1 results today on behalf of our remarkable team and entire community of submitters. This was our seventh major benchmark release of the year, and I’m thrilled by the progress we’ve achieved.
As details on the results show below, you have already proven how fast your systems are on-prem and at the edge in v6.1. Many of you are also shipping these same models via API to millions of users. If that’s your next frontier, bring your v6.1 momentum to MLPerf Endpoints. Rolling submissions to Endpoints open on October 12th! If you’re already running Inference benchmarks, you’re halfway there. We would love to see submitters setting the pace for real-world API inference. Start your MLPerf Endpoints submission here.
Check out the full v6.1 results and details here: MLPerf Inference v6.1 Results Blog
Spread the word! Please like, share, repost: LinkedIn, X / Twitter, Bluesky
Highlights from the latest v6.1 Results:
A record-setting 30 submitting organizations
First Edge Agentic Inference benchmark - peer-reviewed test for multi-turn agentic workloads
First End-to-End RAG benchmark - a full pipeline from query to answer
Largest system ever submitted: 512 accelerators
Two novel heterogeneous systems - cross-vendor and cross-Pacific configurations
Five new processors/accelerators: AMD Ryzen AI Max+ 395, AMD Instinct MI350P, Intel Arc Pro B70, NVIDIA Rubin (Preview), NVIDIA Vera Rubin NVL72 (Preview)
I want to offer particular congratulations and a warm welcome to new submitters: Atlas Inference, Crusoe, Orrick Industries LLC, ScitiX, VibeHPC, and individual contributor Naeem Khoshnevis. Special shoutout to our academic participants—this benchmark is powered by all corners of the community.
Participants and Organizations: AMD, ASUSTeK, Atlas Inference, Cisco, CoreWeave, Crusoe, Dell, Fujitsu Limited, GigaComputing, Google, Hewlett Packard Enterprise, Intel, Inventec Corporation, KRAI, Lambda, MangoBoost, Microsoft Azure, MiTAC, Nebius, NVIDIA, Oracle, Orrick, Quanta Cloud Technology, RedHat, ScitiX, Supermicro, Telecommunications Technology Association, VibeHPC, Wiwynn and individual contributor Naeem Khoshnevis
Partners & Working Group Members: MLPerf has never been a solo sport, and this round is no exception. Here is the tremendous group of people who made it happen: As always, MLPerf is a team activity, and this round in particular has a huge cast of heroes and heroines:
Miro Hodak & Frank Han - co-chairs of the Inference WG
Viraat Chandra (Task force Chair, NVIDIA) and Zhihan Jiang (NVIDIA), for driving the gpt-oss updates
Mingyan Ma (task force Chair, NVIDIA), Shang Wang (NVIDIA), for driving the VLM updates
Ramesh Chukka (Task force Chair, Intel) and Rajesh Poornachandran (Task force Chair, AMD), for creating End to End RAG benchmark.
Palanivel Guruva reddiar (Task force Chair, NVIDIA), Zihao Kong (NVIDIA), and Azeez Ishaqui (Atlas Inference) for creating the agentic edge inference benchmark.
Rashid Kaleem (NVIDIA), Zhihan Jiang (NVIDIA), Miro Hodak (AMD), Frank Han (Dell), Tin-Yin La (NVIDIA), and Rod Shojaei, for helping migrate benchmarks to the API-centric harness for v6.1
Reilly Fairbanks and Scott Wasson, who shepherded such a tremendous release to the finish line
Pablo Gonzales and Rod Shojaei, who led the review process
Lori Blonn and Betsy Masiello, who drove the blog and handled our marketing and communications. in one of the busiest pushes of the year
Kevin Schofeld, who wrote and revised the press release
Pablo Gonzalez Mesa, Tanvi Gour, Arav Agarwal, Anandhu Sooraj, Joe Karmarcik, and Karl Pietri for handling all our IT and infrastructure
Congrats again to everyone—by doing our part to make AI faster, more capable, and more energy efficient, we're helping buyers make informed decisions with confidence.
Excited for what’s next!
Thanks,
David Kanter
Founder, Head of MLPerf
MLCommons