light estimateLast updated 2026-06-19

Lowest-latency real-time / streaming transcription API

For live, low-latency streaming, ElevenLabs Scribe v2 and AssemblyAI Universal-3 Pro both publish around 150 ms — the lowest here — though ElevenLabs runs streaming on a separate realtime mode. Deepgram Nova-3 markets sub-300 ms and is built streaming-first with per-second billing. These numbers mix measured P50 with vendor claims and aren't strictly comparable, so weigh them loosely. A light estimate from public sources, not a first-hand latency benchmark.

DefaultElevenLabs Scribe v2laagste latencyElevenLabs Scribe v2
Provider offerings compared on Price, WER, Langs, Latency and capabilities
OfferingPrice ($/1000 min)WERLangsLatencyCapabilities
ElevenLabs Scribe v2ElevenLabs3.672.2%90150 msdiarizationtimestampsvocab
AssemblyAI Universal-3 ProAssemblyAI3.53.1%99150 msdiarizationtimestampsvocab
Deepgram Nova-3Deepgram4.35.2%50300 msdiarizationtimestampsvocab
Speechmatics EnhancedSpeechmatics6.74%70500 msdiarizationtimestampsvocab
OpenAI gpt-4o-transcribeOpenAI64%
Google Gemini 3 FlashGoogle1.92*2.9%

* token-/credit-priced — the headline understates real per-unit cost, so it is excluded from the cheapest ranking.