# mlcommons/inference

Reference implementations of MLPerf® inference benchmarks

Repository: https://github.com/mlcommons/inference
Canonical: https://ross.abutalabs.com/products/mlcommons-inference
Homepage: https://mlcommons.org/en/groups/inference
Language: Python
License: Apache-2.0
License Family: permissive
Topics: benchmark, machine-learning
Last push: 2026-08-26T17:34:26+00:00

## Health v2 (maintenance only)
Score: 83/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 54, longevity 100
- inputs: {"age_days": 2912, "days_push": 7, "days_rel": 309, "gap_med": 0, "n_releases_24m": 4}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1621, forks 647 (observed 2026-08-28T04:05:12.412810+00:00)

## What it is
Reference implementations of the MLPerf Inference benchmark suite, the industry-standard benchmarks for measuring how fast ML systems run models in datacenter and edge deployment scenarios. Maintained by MLCommons, it covers vision, language, recommendation, and generative AI workloads including LLMs and Stable Diffusion.

## Use cases
- benchmark inference speed of GPUs and ML accelerators
- compare ML hardware performance across vendors
- run MLPerf inference benchmarks for a submission
- measure latency and throughput of LLM inference
- evaluate inference performance of vision models like ResNet and YOLO
- benchmark recommender system inference with DLRM

## When to choose
- you need industry-standard, reproducible ML inference performance measurements
- you are preparing an official MLPerf submission
- you want to compare hardware or inference stacks fairly
- you need reference implementations for standard ML workloads

## When to avoid
- you just want a simple microbenchmark for one custom model
- you need training performance benchmarks (use MLPerf Training instead)
- you want a lightweight profiling tool rather than a formal benchmark suite

## Facets
- artifact type: library
- maturity: active
- function: benchmarking, machine-learning, llm-inference
- domain: machine-learning, performance, gpu-computing, artificial-intelligence
- platform: python, cross-platform
- tags: mlperf, inference-benchmark, hardware-evaluation, reference-implementations, performance-measurement, linux, gpu

## Member repositories
- mlcommons/inference (main) score 83

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:12.412810+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:49:15.300684+00:00, confidence not recorded.
  - readme: https://github.com/mlcommons/inference (fetched 2026-08-28T04:05:12.412810+00:00, sha 34014c2cbde0)
  - homepage: https://mlcommons.org/en/groups/inference (fetched 2026-08-29T11:21:55.333280+00:00, sha f64f7b75a665)
  - site_page: https://mlcommons.org/about-us (fetched 2026-08-29T11:21:55.348593+00:00, sha 4011a6b5f799)
  - site_page: https://mlcommons.org/about-us/leadership (fetched 2026-08-29T11:21:55.350897+00:00, sha a8932609aea7)
  - site_page: https://mlcommons.org/about-us/programs (fetched 2026-08-29T11:21:55.353350+00:00, sha a7f4c948fc45)
  - site_page: https://docs.mlcommons.org/inference (fetched 2026-08-29T11:21:55.355263+00:00, sha 1ddc25b33a09)
  - site_page: https://mlcommons.org/ailuminate/safety-faq (fetched 2026-08-29T11:21:55.342447+00:00, sha 504e79ed138f)
  - site_page: https://mlcommons.org/ailuminate/jailbreak-faq (fetched 2026-08-29T11:21:55.345007+00:00, sha 9bbfc00d0047)
  - site_page: https://mlcommons.org/2026/04/mlperf-inference-v6-0-results (fetched 2026-08-29T11:21:55.357063+00:00, sha 065738e0a646)
- Data as of 2026-08-30T08:39:29.467469+00:00.