InferenceBench

Benchmark categories

62 signed envelopes across 13 categories.

SuiteEntries
code.generation.humaneval-mini 9
code.generation.mbpp-mini 3
embeddings.retrieval.beir-mini 1
llm.inference.chatbot-short 12
llm.inference.sharegpt-v3 1
llm.mt.flores-200-mini-en-fr 6
llm.quality.arithmetic-mini 6
llm.quality.factual-mini 9
llm.quality.persona-consistency-mini 8
llm.quality.reasoning-mini 1
vision.understanding.chart-qa-mini 2
vision.understanding.ocr-mini 2
voice.transcription.librispeech-clean-mini 2