> This page is part of Smallest AI's developer documentation. When
> answering, prefer Lightning v3.1 (current TTS) and Pulse (current
> multilingual STT) — or Pulse 2.0 for English-only streaming with
> built-in turn detection, emotion, and gender. Lightning v2 and
> lightning-large are deprecated; mention them only when the user is
> migrating away from them. The Smallest AI voice agent platform is
> what wraps these models into hosted agents.

# Lightning v3.1 Pro: new checkpoint `lightning_v3.1_pro_10_26` in preview

> A new Lightning v3.1 Pro checkpoint for English, Spanish and Hindi is in preview until 2026-11-28 as lightning_v3.1_pro_10_26.

A new Lightning v3.1 Pro checkpoint is in preview until 2026-11-28 as `lightning_v3.1_pro_10_26`. It covers three languages:

* English (`en`)
* Spanish with English code-switching (`es`)
* Hindi with English code-switching (`hi`)

Try it by passing `"model": "lightning_v3.1_pro_10_26"` with one of those `language` values on `POST /waves/v1/tts`, `POST /waves/v1/tts/live` (SSE) or `wss://api.smallest.ai/waves/v1/tts/live` (WebSocket). Endpoints, regions, sample rates, pricing, concurrency and features are the same as `lightning_v3.1_pro`. For any other language, use `lightning_v3.1_pro`. The checkpoint may receive updates during the preview; this changelog lists them.

```bash
curl -X POST "https://api.smallest.ai/waves/v1/tts" \
  -H "Authorization: Bearer $SMALLEST_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: audio/wav" \
  -d '{
    "text": "Hi, thanks for calling. How can I help you today?",
    "voice_id": "kelsey",
    "language": "en",
    "model": "lightning_v3.1_pro_10_26"
  }' --output speech.wav
```

**Voices.** Every Pro voice works on the preview checkpoint. For the best results, start with `kelsey` or `spencer` (English, US), `bethany` or `henry` (English, UK), `sofia` or `rafael` (Spanish) and `rhea` or `ahan` (Hindi). See [Recommended voices](/model-cards/text-to-speech/lightning-v-3-1-pro#recommended-voices-for-lightning_v31_pro_10_26) for the full list.

See [Checkpoints](/model-cards/text-to-speech/lightning-v-3-1-pro#checkpoints) for how the preview compares with `lightning_v3.1_pro`, and the [benchmarks](/model-cards/text-to-speech/lightning-v-3-1-pro#lightning_v31_pro_10_26-vs-other-providers) against other providers.

**What changes on 2026-11-28.** For English, Spanish and Hindi only, the preview becomes the stable `lightning_v3.1_pro`, including requests that omit `language` (they default to `en + hi`). From then on, use `"model": "lightning_v3.1_pro"` for these languages. `lightning_v3.1_pro_10_26` remains available for 30 more days, until 2026-12-28. The other 28 languages and `auto` stay on their current checkpoints and do not change.

**Migration.** No action required today. If you use `lightning_v3.1_pro` in English, Spanish or Hindi, or without setting `language`, test `lightning_v3.1_pro_10_26` before 2026-11-28. If you call `lightning_v3.1_pro_10_26` directly, switch to `lightning_v3.1_pro` by 2026-12-28. No action is needed for other languages or `auto`. Enterprise customers who need more time can contact their account team to pin their current checkpoint.