# Kalpa TTS Beta v0.1 Benchmark Scores & Performance

> Kalpa TTS Beta v0.1 has Audio Realism results. Voice evidence is display-only and does not produce a text-model score.

Canonical page: https://benchlm.ai/models/kalpa-tts-beta-v0-1

Last updated: October 17, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | Kalpa Labs |
| Source Type | Proprietary |
| Reasoning Type | Non-Reasoning |
| Context Window | N/A |
| Official model card | [Kalpa model documentation](https://docs.kalpalabs.ai/models) |
| Overall Score | Not scored (voice evidence only) |
| Overall Rank | Unranked |

## Family & Coverage

- Family: Kalpa TTS
- Variant: beta-v0-1 (Beta v0.1)
- Benchmarks covered: 0 of 639
- Coverage note: Speech and voice results appear below; no weighted text-model benchmark rows are stored.

## Speech and voice evidence

[Audio Realism Benchmark](/voice-benchmarks/audio-realism-benchmark)

The leaderboard measures perceived human-likeness for the selected American-English voices and prompt mix. It does not measure factual accuracy, task completion, full-duplex interaction, or multilingual quality.

| Metric | Result |
| --- | --- |
| Owner rank | #3 of 29 |
| Elo-style rating | 1,364 |
| BT uncertainty | Not published |
| Win rate | 74.4% |
| Battles | 540 |
| Avg. generation | 2.70s |

402 wins · 138 losses · owner snapshot 2026-10-11

- [Evaluation source](https://www.designarena.ai/methodology/audio-realism-benchmark)

Available in early access through the Kalpa Speech API as `kalpa-tts-beta-v0.1`; the provider describes it as the default English conversational speech model for dialogue and text-to-speech.