# A Preliminary Exploration with GPT-4o Voice Mode

arXiv: https://arxiv.org/abs/2502.09940 · HTML: https://mlsanity.com/paper/2502.09940/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- Placed in Video understanding with MLLMs, one of 16 directions in Multimodal LLMs / vision-language models.
- First posted on arXiv 14 Feb 2025.

## Placed in

- Multimodal LLMs / vision-language models / [Video understanding with MLLMs](https://mlsanity.com/mllm/video-understanding-with-mllms/)

## Built on by (most cited first, 16 of 16)

- [SAKURA: On the Multi-hop Reasoning of Large Audio-Language Models Based on Speech and Audio Information](https://mlsanity.com/paper/2505.13237/) (2025)
- [Towards Holistic Evaluation of Large Audio-Language Models: A Comprehensive Survey](https://mlsanity.com/paper/2505.15957/) (2025)
- [AudioLens: A Closer Look at Auditory Attribute Perception of Large Audio-Language Models](https://mlsanity.com/paper/2506.05140/) (2025)
- [All That Glitters Is Not Audio: Rethinking Text Priors and Audio Reliance in Audio-Language Evaluation](https://mlsanity.com/paper/2604.24401/) (2026)
- [SAKE: Towards Editing Auditory Attribute Knowledge of Large Audio-Language Models](https://mlsanity.com/paper/2510.16917/) (2025)
- [Hierarchical Acoustic-Semantic Modeling: Modality Separation and Semantic Coherence for Full-Duplex SLMs](https://mlsanity.com/paper/2607.06540/) (2026)
- [When Silence Matters: The Impact of Irrelevant Audio on Text Reasoning in Large Audio-Language Models](https://mlsanity.com/paper/2510.00626/) (2025)
- [Investigating Safety Vulnerabilities of Large Audio-Language Models Under Speaker Emotional Variations](https://mlsanity.com/paper/2510.16893/) (2025)
- [ASPIRin: Action Space Projection for Interactivity-Optimized Reinforcement Learning in Full-Duplex Speech Language Models](https://mlsanity.com/paper/2604.10065/) (2026)
- [MUGEN: Evaluating and Improving Multi-audio Understanding of Large Audio-Language Models](https://mlsanity.com/paper/2603.09714/) (2026)
- [Nudging Hidden States: Training-Free Model Steering for Chain-of-Thought Reasoning in Large Audio-Language Models](https://mlsanity.com/paper/2603.14636/) (2026)
- [Identifying and typifying demographic unfairness in phoneme-level embeddings of self-supervised speech recognition models](https://mlsanity.com/paper/2604.22631/) (2026)
- [Towards Generalized Source Tracing for Codec-Based Deepfake Speech](https://mlsanity.com/paper/2506.07294/) (2025)
- [Generative Multi-modal Feedback for Singing Voice Synthesis Evaluation](https://mlsanity.com/paper/2512.02523/) (2025)
- [Style Amnesia: Investigating Speaking Style Degradation and Mitigation in Multi-Turn Spoken Language Models](https://mlsanity.com/paper/2512.23578/) (2025)
- [Encoder-Side Neuron Identification and Amplification for Acoustic Perception in Large Audio-Language Models](https://mlsanity.com/paper/2607.11801/) (2026)

## Abstract (arXiv)

With the rise of multimodal large language models, GPT-4o stands out as a pioneering model, driving us to evaluate its capabilities. This report assesses GPT-4o across various tasks to analyze its audio processing and reasoning abilities. We find that GPT-4o exhibits strong knowledge in audio, speech, and music understanding, performing well in tasks like intent classification, spoken command classification, semantic and grammatical reasoning., multilingual speech recognition, and singing analysis. It also shows greater robustness against hallucinations than other large audio-language models (LALMs). However, it struggles with tasks such as audio duration prediction and instrument classification. Additionally, GPT-4o's safety mechanisms cause it to decline tasks like speaker identification, age classification, MOS prediction, and audio deepfake detection. Notably, the model exhibits a significantly different refusal rate when responding to speaker verification tasks on different datasets. This is likely due to variations in the accompanying instructions or the quality of the input audio, suggesting the sensitivity of its built-in safeguards. Finally, we acknowledge that model performance varies with evaluation protocols. This report only serves as a preliminary exploration of the current state of LALMs.