Skip to main content

audio model comparison

Gemini 3.8 Flash-Lite TTS vs Scribe v2 Realtime

Gemini 3.8 Flash-Lite TTS and Scribe v2 Realtime are both current audio models, from Google and ElevenLabs. Scribe v2 Realtime is 28% cheaper per audio minute than Gemini 3.8 Flash-Lite TTS, $0.0065 vs $0.009.

Data verified October 13, 2026 · Dataset 2026-10-13

Share this

Side by side

Standard-tier prices in effect today. The better value in each row is highlighted.

FactGemini 3.8 Flash-Lite TTSScribe v2 Realtime
Price per audio minute$0.009$0.0065 (better)
Artificial Analysis Speech Arena1246 Elo (#5 of 11)Not measured
Artificial Analysis AA-WER StreamingNot measured3.59 % WER (#2 of 2)
ProviderGoogleElevenLabs
Release dateSeptember 22, 2026November 11, 2025
AccessAPI onlyAPI only
Context windowNot reportedNot reported
Max outputNot reportedNot reported
Inputtextaudio
Outputaudiotext

Scores:Artificial Analysis · 2026-10-13Artificial Analysis · 2026-10-13

When to choose each

Gemini 3.8 Flash-Lite TTS

Best for

  • High-volume speech generation such as notifications, IVR prompts and accessibility audio

Watch out for

  • Google doubles the rates on 2027-01-01, to $1.00 input and $12 output per 1M tokens (about $0.018 per minute)
  • No independent speech-quality benchmark is recorded here
Full Gemini 3.8 Flash-Lite TTS profile

Scribe v2 Realtime

Best for

  • Live transcription

Watch out for

  • $0.39 per hour of audio, about 1.8 times batch Scribe v2
Full Scribe v2 Realtime profile

Gemini 3.8 Flash-Lite TTS vs Scribe v2 Realtime FAQ

Which is cheaper, Gemini 3.8 Flash-Lite TTS or Scribe v2 Realtime?

Scribe v2 Realtime is 28% cheaper per audio minute than Gemini 3.8 Flash-Lite TTS, $0.0065 vs $0.009.

Production evaluation

Choosing between Gemini 3.8 Flash-Lite TTS and Scribe v2 Realtime?

Public benchmarks narrow the shortlist. We run both on your own tasks, data and cost limits, then build the routing and review controls to run the winner in production.

  • A head-to-head evaluation on your real workload
  • Routing between models by cost and difficulty