ResearchDatasetsWebLLMFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsLLM InferenceWebLLM macOS Apple M2 16GB unified memory46.4Decode Throughput (tok/s)6LLM InferenceWebLLM Windows 11 RTX PRO 2000 Blackwell 8GB51.1Decode Throughput (tok/s)4