Huawei Ascend 950PR microbenchmarksNew
Cycle-level core memory-access microbenchmarks (beta)
Huawei Ascend 950PR microbenchmarksNew
950PR GEMM FLOPs and max FLOPS utilization
Compare realistic agentic inference performance
Long-context, multi-turn inference performance on Google TPUv7 Ironwood, OpenAI Jalapeño, MI355X, GB300 NVL72, GB200 NVL72, B200, H200, H100 and RTX Pro. TPUv8i, SambaNova SN50, Rubin NVL72 and MI455X UALoE72 are next.
Open-source continuous AI accelerator system benchmark trusted by gigawatt token factories
Every result comes from a public GitHub Actions run
Every data point on the dashboard is produced by a public GitHub Actions workflow run. The recipe lives in the repo, the run executes on the actual target hardware, and the full logs and artifacts are publicly viewable. Click any point on a chart to jump straight to the run that produced it. All reproducible, auditable, and open source.
1,000+ new benchmark datapoints added per week on average. Browse every new model, chip, framework, and configuration as it lands.
- Public Actions runs
- Every benchmark executes on GitHub Actions with full logs visible while the run is in progress.
- Open recipes
- Every model, framework, precision, and parallelism setting is committed to the public repo as a shell script.