Model Performance Benchmark

Speech Model Performance

Track chip performance on ASR, TTS, speech enhancement, and latency-sensitive voice workloads across cloud and edge environments.

Benchmark Focus

ASR, TTS, streaming speech, and audio inference workloads

Comparison Coverage

4 models / 6 profiles

Choose one model and one shared test condition to compare chips side by side

Tracked Models

Published speech models with active benchmark records

Comparable Profiles

Core filters resolve into one shared benchmark profile before table rendering

Chips Covered

Cloud and edge accelerators with published benchmark results

Comparison controls

Pick one model, refine high-frequency conditions in the core area, then use advanced filters only when you need deeper control.

Matching profiles 1

Model

Core Conditions

TaskModeAudio DurationSample RatePrecisionQuantizationFramework

Advanced Conditions

Expose lower-frequency benchmark knobs without crowding the default comparison flow.

SegmentManufacturer

Comparing 3 chip results

Selected model

MMS English ASR

INT8 ASR · Meta

Primary metric: throughput (x realtime)

Task speech-recognitionFramework ONNX RuntimeRuntime Optimized EPBatch 8Precision INT8Quant INT8Variant INT8 ASRSubtype asrAudio 15sSample Rate 16000HzStreaming NoLanguage English

Chip	Segment	Manufacturer	throughput	Throughput	Latency P50	Latency P99	Power	Memory	Source
NVIDIA H200 SXM 989 TFLOPS	Cloud	NVIDIA	34.1 x realtime	34.1 x realtime	—	—	500 W	14 GB	View Source
Unknown chip —	Edge	—	6.4 x realtime	6.4 x realtime	—	—	16 W	3.4 GB	View Source
Unknown chip —	Edge	—	4.7 x realtime	4.7 x realtime	—	—	17 W	2.2 GB	View Source

Loading benchmark controls...

Browse Cloud Chips Browse Edge Chips