Positioning and release
Gemini 3.8 Flash-Lite TTS is a fast, cost-efficient text-to-speech model for high-throughput production and low-latency voice applications. Google recommends it as a replacement for gemini-3.1-flash-tts-preview. The Gemini API release notes list its general availability on September 22, 2026; Google's blog and Firebase model list give September 23, 2026. (aistudio.google.com)
Capabilities and uses
- Accepts text and generates audio; supports 101 languages with automatic language detection.
- Designed for bulk speech production, read-aloud features, conversational voice-agent cascades, and everyday single-speaker generation.
- Supports prebuilt voices, the extended voice library, custom voice design, voice replication and voice remixing. Voice replication requires verification of the voice owner's consent.
- Shares its API schema and structured prompting format with
gemini-3.8-flash-tts. Unary requests return WAV audio by default. (aistudio.google.com)
Limits and considerations
- Input limit: 8,192 tokens. Output limit: 16,384 tokens under the Gemini API serving limit.
- Does not support the Live API, context caching, function calling, code execution, structured output, thinking, search grounding or image generation. Use in a voice-agent cascade does not imply native Live API support.
- Input text is treated as a transcript; delivery style and speaker labels should be supplied as structured speech metadata rather than ordinary inline instructions. Google's audio model card notes possible hallucinations, slow responses or timeouts. (aistudio.google.com)
Gemini Developer API pricing
Prices below are in USD per 1 million tokens, for this model and the specified service mode. Pricing checked November 16, 2026. The official page lists free-tier input and output for Standard and Priority, subject to applicable availability and limits; Batch and Flex have no free tier. (aistudio.google.com)
| Paid service mode | Text input through Dec 31, 2026 | Audio output through Dec 31, 2026 | Text input from Jan 1, 2027 | Audio output from Jan 1, 2027 |
|---|---|---|---|---|
| Standard | $0.50 | $6.00 | $1.00 | $12.00 |
| Batch | $0.25 | $3.00 | $0.50 | $6.00 |
| Flex | $0.25 | $3.00 | $0.50 | $6.00 |
| Priority | $0.90 | $10.80 | $1.80 | $21.60 |
Google states that audio corresponds to 25 tokens per second. These are Gemini Developer API rates, not a claim about pricing on other platforms. (aistudio.google.com)