Get startedLeaderboardDecode calculatorModelsReportsHardwareBenchmarksMarketplaceRentalsProAPI Docs
Language
Actual Computer — Every computer, one endpoint

Observed performance

Run snapshot

Current results after the filters below are applied.
Fastest run
Median speed
Memory ceiling
Slowest run
Measured runs
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 — 3,662.0 tok/s on H100 · 80 GB | Localmaxxing