GOESB

GOESB — Global Open Edge Speech Benchmark · pronounced “goose-bee”

Leaderboard

Objective, reproducible benchmarks for local (edge) speech AI. Every result here re-verified its hash and signature before being accepted — see why you can trust the numbers. New here? Read the docs.

This is every verified result, filterable and sortable. Looking for a single best recommendation for your constraints instead? Try the homepage →

Lowest energyFastest (lowest RTF)Most accurate (lowest WER)
StreamingStreaming results only — real-time factor doesn't apply, so batch results never appear here.
AnyBatchStreamingConcurrency

Sorted by most recently submitted · 1 result · click a metric below to sort by it

ProfileRuntime / ModelHardwarePackMetricsSubmittedDetails
whispercpp-large-v3-es-streamingwhisper-cpp / whisper-large-v3metalapple-m1-profleurs-es
wer: 0.0399 ratioram_mb: 4063.33 MBcpu_pct: 8.20 %real_time_factor: 1.07 ratioupdate_frequency: 0.9534 Hzpartial_stability: 0.9450 ratiofirst_final_latency: 5233.96 msend_of_speech_latency: 2160.05 msfirst_partial_latency: 1734.84 msstreaming_responsiveness: 0.5007 index
2026-08-03T14:27:31Z
details
Profile
whispercpp-large-v3-es-streaming@1.0.0
Pack
fleurs-es@1.0.0
Repeats
2
Backend
metal
chunk_ms
1000
threads
4

Constant across this view: chunk_ms 1000 · threads 4

* profile default