# Speech-to-Speech Question Answering on Llama Questions Source: [Speech-to-Speech Question Answering on Llama Questions]() Collection index: [llms.txt]() Data updated: 2026-04-21T12:43:57.695Z Evaluation of model performance on the Llama Questions dataset for Speech-to-Speech Question Answering, measured using Accuracy. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Speech-to-Speech Question Answering | | Dataset (including any stated split) | Llama Questions | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 27 of 27 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | FreezeEmpath | 79.33 | Not recorded | 69e704e17d0172140ec7f0c4 | | Kimi-Audio | 79 | Not recorded | 69e704e17d0172140ec7f0c4 | | VocalNet | 78.6 | Not recorded | 69b0f7312d603e6de3e46000 | | Fun-Audio-Chat-8B | 77.76 | Not recorded | 69e704e17d0172140ec7f0c4 | | Qwen2.5-Omni-7B | 77.33 | Not recorded | 69b0f7312d603e6de3e46000 | | AudioKV | 77.3 | Not recorded | 69d732b37d0172140ec76c87 | | PyramidKV | 76.7 | Not recorded | 69d732b37d0172140ec76c87 | | SnapKV | 76 | Not recorded | 69d732b37d0172140ec76c87 | | AdaKV | 76 | Not recorded | 69d732b37d0172140ec76c87 | | AudioKV | 75 | Not recorded | 69d732b37d0172140ec76c87 | | AdaKV | 74.7 | Not recorded | 69d732b37d0172140ec76c87 | | PyramidKV | 74.7 | Not recorded | 69d732b37d0172140ec76c87 | | FreezeEmpath | 74.67 | Not recorded | 69e704e17d0172140ec7f0c4 | | LLaMA-Omni2-7B | 73.67 | Not recorded | 69e704e17d0172140ec7f0c4 | | SPEECH-OMNI-LITE | 73.33 | Not recorded | 69b0f7312d603e6de3e46000 | | LLaMA-Omni2 | 72 | Not recorded | 69b0f7312d603e6de3e46000 | | SnapKV | 72 | Not recorded | 69d732b37d0172140ec76c87 | | Fun-Audio-Chat-8B | 72 | Not recorded | 69e704e17d0172140ec7f0c4 | | OpenS2S | 70.67 | Not recorded | 69e704e17d0172140ec7f0c4 | | Step-Audio2-Mini | 66.67 | Not recorded | 69e704e17d0172140ec7f0c4 | | LLaMA-Omni2-7B | 66.67 | Not recorded | 69e704e17d0172140ec7f0c4 | | Baichuan-Omni-1.5 | 65.67 | Not recorded | 69b0f7312d603e6de3e46000 | | Kimi-Audio | 64.67 | Not recorded | 69e704e17d0172140ec7f0c4 | | Step-Audio2-Mini | 64.33 | Not recorded | 69e704e17d0172140ec7f0c4 | | GLM-4-Voice | 63.67 | Not recorded | 69b0f7312d603e6de3e46000 | | OpenS2S | 59 | Not recorded | 69e704e17d0172140ec7f0c4 | | Freeze-Omni | 58.67 | Not recorded | 69b0f7312d603e6de3e46000 | ### FreezeEmpath Accuracy: 79.33. Result ID: 69e770f947f34ce924cf21d2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2T"}. This record does not assert independent reproduction. ### Kimi-Audio Accuracy: 79. Result ID: 69e770f947f34ce924cf21cf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2T"}. This record does not assert independent reproduction. ### VocalNet Accuracy: 78.6. Result ID: 69b131466dbb7075b95bce02. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"8B"}. This record does not assert independent reproduction. ### Fun-Audio-Chat-8B Accuracy: 77.76. Result ID: 69e770f947f34ce924cf21ce. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2T"}. This record does not assert independent reproduction. ### Qwen2.5-Omni-7B Accuracy: 77.33. Result ID: 69b131466dbb7075b95bce06. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"7B"}. This record does not assert independent reproduction. ### AudioKV Accuracy: 77.3. Result ID: 69d83ec647f34ce924c03621. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-Omni-7B","KV cache budget":"40%","max\_gen\_tokens":"64","Judge":"Gemini 2.5 Flash"}. This record does not assert independent reproduction. ### PyramidKV Accuracy: 76.7. Result ID: 69d83ec647f34ce924c03620. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-Omni-7B","KV cache budget":"40%","max\_gen\_tokens":"64","Judge":"Gemini 2.5 Flash"}. This record does not assert independent reproduction. ### SnapKV Accuracy: 76. Result ID: 69d83ec647f34ce924c0361e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"Qwen2.5-Omni-7B","KV cache budget":"40%","max\_gen\_tokens":"64","Judge":"Gemini 2.5 Flash"}. This record does not assert independent reproduction. ### AdaKV Accuracy: 76. Result ID: 69d83ec647f34ce924c0361f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-Omni-7B","KV cache budget":"40%","max\_gen\_tokens":"64","Judge":"Gemini 2.5 Flash"}. This record does not assert independent reproduction. ### AudioKV Accuracy: 75. Result ID: 69d83ec647f34ce924c0361d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-Omni-3B","KV cache budget":"40%","max\_gen\_tokens":"64","Judge":"Gemini 2.5 Flash"}. This record does not assert independent reproduction. ### AdaKV Accuracy: 74.7. Result ID: 69d83ec647f34ce924c0361b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-Omni-3B","KV cache budget":"40%","max\_gen\_tokens":"64","Judge":"Gemini 2.5 Flash"}. This record does not assert independent reproduction. ### PyramidKV Accuracy: 74.7. Result ID: 69d83ec647f34ce924c0361c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-Omni-3B","KV cache budget":"40%","max\_gen\_tokens":"64","Judge":"Gemini 2.5 Flash"}. This record does not assert independent reproduction. ### FreezeEmpath Accuracy: 74.67. Result ID: 69e770f947f34ce924cf21d8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2S"}. This record does not assert independent reproduction. ### LLaMA-Omni2-7B Accuracy: 73.67. Result ID: 69e770f947f34ce924cf21d1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2T"}. This record does not assert independent reproduction. ### SPEECH-OMNI-LITE Accuracy: 73.33. Result ID: 69b131466dbb7075b95bce07. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"8B"}. This record does not assert independent reproduction. ### LLaMA-Omni2 Accuracy: 72. Result ID: 69b131466dbb7075b95bce03. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"7B"}. This record does not assert independent reproduction. ### SnapKV Accuracy: 72. Result ID: 69d83ec647f34ce924c0361a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"Qwen2.5-Omni-3B","KV cache budget":"40%","max\_gen\_tokens":"64","Judge":"Gemini 2.5 Flash"}. This record does not assert independent reproduction. ### Fun-Audio-Chat-8B Accuracy: 72. Result ID: 69e770f947f34ce924cf21d4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2S"}. This record does not assert independent reproduction. ### OpenS2S Accuracy: 70.67. Result ID: 69e770f947f34ce924cf21d0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2T"}. This record does not assert independent reproduction. ### Step-Audio2-Mini Accuracy: 66.67. Result ID: 69e770f947f34ce924cf21cd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2T"}. This record does not assert independent reproduction. ### LLaMA-Omni2-7B Accuracy: 66.67. Result ID: 69e770f947f34ce924cf21d7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2S"}. This record does not assert independent reproduction. ### Baichuan-Omni-1.5 Accuracy: 65.67. Result ID: 69b131466dbb7075b95bce05. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"7B"}. This record does not assert independent reproduction. ### Kimi-Audio Accuracy: 64.67. Result ID: 69e770f947f34ce924cf21d5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2S"}. This record does not assert independent reproduction. ### Step-Audio2-Mini Accuracy: 64.33. Result ID: 69e770f947f34ce924cf21d3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2S"}. This record does not assert independent reproduction. ### GLM-4-Voice Accuracy: 63.67. Result ID: 69b131466dbb7075b95bce01. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"9B"}. This record does not assert independent reproduction. ### OpenS2S Accuracy: 59. Result ID: 69e770f947f34ce924cf21d6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"S2S"}. This record does not assert independent reproduction. ### Freeze-Omni Accuracy: 58.67. Result ID: 69b131466dbb7075b95bce04. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Parameters":"7B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.