ServiceNow releases open Apriel Voice 27B, closing the gap with proprietary voice models
jacek2023 · reddit · 2026-10-30
ServiceNow AI released Apriel Voice 27B Preview, an open speech model built on a Gemma 4 26B-A4B backbone with a Whisper large-v3-turbo speech encoder, supporting transcription, speech translation, spoken reasoning, and tool calling.
Benchmarks: 51.27% EVA-A, 74.13% EVA-X, 65.51% τ-Voice task success at 1.91s mean latency; 2.23% WER on LibriSpeech test-clean. Notably it beats Gemini-3.8 Live Extended Thinking on EVA-X (74.13 vs 73.5) and crushes other open alternatives like UltraVox (1.91% EVA-A).
Long audio: recordings auto-chunked at 30s; tested good up to 30 minutes, with thinking mode recommended beyond. The checkpoint outputs text and tool calls rather than native full-duplex speech — pair with an external synthesizer for live voice apps. Demo, blog, and technical report available.
More from Models
- LLMs finally write more secure code: Veracode benchmark average jumps 56% to 62% — WeldPond · 2026-10-30
- Pedro Domingos: Claude is 'pretty useless' for code never written before — pmddomingos · 2026-10-30
- AI Code Security Finally Improves: Veracode Benchmark Pass Rate Jumps From 56% to 62% — WeldPond · 2026-10-30
- Anthropic exec confirms $500 plan can nearly saturate most bug hunters' agent usage — rez0__ · 2026-10-30
- 40 days of coding-agent telemetry show a provider silently swapped model weights — Jipok_ · 2026-10-30
- Local Gemma 'completely good enough' for non-coding tasks, designer finds — michalmalewicz · 2026-10-30