arXiv cs.CL· Aryan Vijay Bhosale, Harshit Rajgarhia, Abhishek Mukherji, Dinesh Manocha·· 13 小时前AI 评分38
TORUS：面向统一音频模型的闭环基准发布
TORUS: A Unified Closed-Loop Benchmark for Unified Audio Models
AI 导读
TORUS 是一个音频原生统一基准，用 1,053 道题（117 项测试）覆盖 21 个任务、8 种音频场景设置和全部 3 类音频，闭环评估统一音频模型的理解-生成自洽性（UGC_OWN）。测试显示最佳统一模型只能答对自身生成音频的 52.4% 问题，gpt-audio 为 53.6%。研究据此把理解-生成一致性列为未来统一音频系统的关键测试项。
正文 · 原文
Abstract:Unified audio models capable of audio understanding, audio generation and, increasingly, audio editing are proliferating rapidly. However, each capability is evaluated in isolation on specialized benchmarks and a basic question remains unanswered: can these models understand the audio they generate/edit i.e. do they exhibit audio understanding-generation self-coherence (UGC_OWN)? We introduce TORUS, an audio-native unified benchmark for holistic, closed-loop evaluation of unified audio models comprising 1,053 questions (117 tests) spanning 21 tasks and 8 audio domain settings across all 3 audio types. We evaluate unified models and cascades of specialized models across objective and subjective metrics. Best unified models are able to answer only 52.4% questions about their generations, while gpt-audio can answer 53.6% questions. We hypothesize that a truly unified audio intelligence should be highly self-coherent and thus position understanding-generation coherence as an essential test for future unified audio systems. Further details are on our project page: this https URL
| Subjects: | Sound (cs.SD); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2607.28896 [cs.SD] |
| (or arXiv:2607.28896v2 [cs.SD] for this version) | |
| https://doi.org/10.48550/arXiv.2607.28896 arXiv-issued DOI via DataCite |
Submission history
From: Aryan Vijay Bhosale [view email]
[v1]
Thu, 30 Jul 2026 23:28:11 UTC (4,643 KB)
[v2]
Wed, 21 Oct 2026 17:45:20 UTC (4,825 KB)
来源：arXiv cs.CL · arxiv.org