# SeedRealtime launch notes: Voice Benchmark Profile

> ByteDance Seed’s August 5, 2026 announcement describes an end-to-end audio-visual full-duplex model and reports a halving of conversational pacing issues versus cascaded pipelines.

- Scope: Relative conversational-pacing claim versus cascaded ASR–LLM–TTS systems and architecture disclosures
- Measurement lanes: Conversation dynamics
- Primary metric: Documented capabilities (no scored benchmark)
- Available evidence: Relative claim only
- Owner: ByteDance Seed
- Source snapshots refreshed: 2026-09-11

## Interpretation limit

Provider claim without a published benchmark table or competitor values.

## Primary sources

- [Owner page](https://seed.bytedance.com/en/SeedRealtime)

Canonical page: https://benchlm.ai/voice-benchmarks/seedrealtime-launch-evaluation