Open Brain
VercelEngineering PostOfficial Source

Gemini 3.8 text-to-speech models now available on AI Gateway

Vercel AI Gateway now exposes two Gemini 3.8 TTS models for multilingual speech, separating high-volume generation from prompt-driven voice design.

Vercel · Sep 23, 2026
Open Source Open MarkdownOpen JSON
Source Summary

AI Gateway now offers **Gemini 3.8 Flash-Lite TTS** and **Gemini 3.8 Flash TTS**. Both generate speech in **more than 100 languages** and support long narration and two-speaker dialogue.

Practical Implication

Use Flash-Lite for high-volume output with tone, pacing, and line-level controls. Choose Flash when an agent needs prompt-driven voice or character design, including acting cues, accents, and reactions.

Agent-Ready Context
AI Gateway now offers **Gemini 3.8 Flash-Lite TTS** and **Gemini 3.8 Flash TTS**. Both generate speech in **more than 100 languages** and support long narration and two-speaker dialogue.

Use Flash-Lite for high-volume output with tone, pacing, and line-level controls. Choose Flash when an agent needs prompt-driven voice or character design, including acting cues, accents, and reactions.

The material describes capabilities and gateway features such as routing, request-level usage tracking, and provider keys, but provides no pricing, latency, quality evaluation, or language-by-language results.
Connected Context · Feed7 Judgment

This creates a concrete intra-family routing decision for speech: Flash-Lite for controlled volume production and Flash for prompt-designed voices and performances. It broadens the Gateway’s audio surface to long narration and two-speaker output across 100-plus languages, but does not establish either model as a default because price, latency, quality, and language-level performance remain unmeasured.

Context Map
infraaudio#generative-media#gateways#model-selection
Uncertainty
The material describes capabilities and gateway features such as routing, request-level usage tracking, and provider keys, but provides no pricing, latency, quality evaluation, or language-by-language results.