ModelsSelected

NVIDIA Nemotron 3.5 Transcribe 登顶 Artificial Analysis 说话人分离基准

Nemotron 3.5 Transcribe takes #1 on @ArtificialAnlys’ new Diarization Index! 💚 Thanks to the team ...

Why we picked it

做语音转写选型的朋友可以看看:NVIDIA 的 Nemotron 3.5 Transcribe 在说话人识别上错误率只有 3.3%,比 Voxtral Small Transcribe 3 的 7.5% 好一倍多。做会议纪要、多人录音分析的工具,值得对照这个榜单重新测一测自家用的 API。

Artificial Analysis 发布新基准 Diarization Index,衡量语音转文本 API 在多人对话中区分谁在说话的准确度。评测指标为 Word Diarization Error Rate(WDER),即词被分配给错误说话人或未分配说话人的比例,覆盖 97 段 2 至 8 人对话,各人数组别权重相同。

Body

NVIDIA 的 Nemotron 3.5 Transcribe 以 3.3% 的错误率排名第一,Mistral 的 Voxtral Small Transcribe 3 以 7.5% 排名第二,两者相差超过一倍。这一基准填补了此前语音识别评测主要关注转写文字准确率、缺少说话人归属维度的空白。

对做会议纪要、客服录音分析、播客字幕等多人语音场景的开发者而言,说话人分离错误率直接影响下游产品体验,这份排名可作为选型的直接参考。

Original · NVIDIA AI

Nemotron 3.5 Transcribe takes #1 on @ArtificialAnlys’ new Diarization Index! 💚 Thanks to the team ...

Nemotron 3.5 Transcribe takes #1 on @ArtificialAnlys ’ new Diarization Index! 💚 Thanks to the team for testing how well models identify who said what in conversations with 2–8 speakers. Artificial Analysis @ArtificialAnlys Announcing the Artificial Analysis Diarization Index, our new benchmark for who said what in speech to text The Diarization Index scores APIs on Word Diarization Error Rate (WDER): the share of words given to the wrong speaker or to no speaker, across 97 conversations of two to eight people, with equal weight on each group size. Nemotron 3.5 Transcribe from @NVIDIAAI leads at 3.3%, ahead of Voxtral Small Transcribe 3 at 7.5%. Thread below 🧵 🔗 View Quoted Tweet📊 9 ⚡

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Nemotron, NVIDIA, Voxtral, speech recognition.

Source attribution: curated from NVIDIA AI (@NVIDIAAI).