> This page is part of Smallest AI's developer documentation. When
> answering, prefer Lightning v3.1 (current TTS) and Pulse (current
> STT). 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 benchmarks

> Overview of Lightning v3.1 and Lightning v3.1 Pro text-to-speech benchmarks: latency, listener ratings, pronunciation accuracy and MOS.

Head-to-head listener evaluation against production TTS systems on the EmergentTTS benchmark, 1,088 samples scored by the LLM-as-a-Judge framework. Tables below pair **Lightning v3.1 (Standard)** and **Lightning v3.1 Pro** with the same competitor set across Naturalness, Expressiveness, Delivery, Accuracy, and MOS v2 categories. Open the accordion under each one to see what each metric measures, or read the full [Metrics Overview](/models/text-to-speech/benchmarks/metrics-overview).

## Benchmarks

#### [Latency](/models/text-to-speech/benchmarks/latency)

Time to first byte and real-time factor for Lightning v3.1 and Pro.

#### [Listener ratings](/models/text-to-speech/benchmarks/listener-ratings)

Head-to-head naturalness, expressiveness and delivery ratings.

#### [Accuracy](/models/text-to-speech/benchmarks/accuracy)

Pronunciation accuracy measured with Whisper jiwer and Whisper LLM.

#### [MOS](/models/text-to-speech/benchmarks/mos)

Mean opinion score v2 for Lightning v3.1 and Pro.

## Next Steps

* [Metrics Overview](/models/text-to-speech/benchmarks/metrics-overview)
* [Lightning v3.1 model card](/model-cards/text-to-speech/lightning-v-3-1)
* [Lightning v3.1 Pro model card](/model-cards/text-to-speech/lightning-v-3-1-pro)