# 3D Visual Question Answering on ScanQA (EM@1, EM@10) Source: [3D Visual Question Answering on ScanQA (EM@1, EM@10)]() Collection index: [llms.txt]() Data updated: 2026-05-07T09:47:18.854Z Evaluation of model performance on the ScanQA dataset for 3D Visual Question Answering, specifically reporting Exact Match metrics at thresholds 1 and 10. 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 | 3D Visual Question Answering | | Dataset (including any stated split) | ScanQA | | Metric | EM@1 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 17 of 17 | | 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 - [EM@1](): Higher is better; unit: PERCENT. - [EM@10](): Higher is better; unit: PERCENT. ## Recorded results | Method | EM@1 | Result date | Report paper | | --- | --- | --- | --- | | Video-3D LLM | 30.1 | Not recorded | 691ff208d6b0dd334292e672 | | LLaVA-3D | 27 | Not recorded | 691ff208d6b0dd334292e672 | | LEO | 24.5 | Not recorded | 691ff208d6b0dd334292e672 | | Qwen2.5-VL-7B | 23.7 | Not recorded | 691ff208d6b0dd334292e672 | | UniScene3D | 23.2 | Not recorded | 69d33e187d0172140ec7377f | | SceneVerse | 22.7 | Not recorded | 69d33e187d0172140ec7377f | | SceneVerse | 22.7 | Not recorded | 691ff208d6b0dd334292e672 | | 3D-VisTA | 22.4 | Not recorded | 69d33e187d0172140ec7377f | | SplatTalk | 22.4 | Not recorded | 691ff208d6b0dd334292e672 | | 3D-ViSTA | 22.4 | Not recorded | 691ff208d6b0dd334292e672 | | POMA-3D | 22.3 | Not recorded | 69d33e187d0172140ec7377f | | POMA-3D\_spec | 22.3 | Not recorded | 691ff208d6b0dd334292e672 | | POMA-3D\_llm | 21.3 | Not recorded | 691ff208d6b0dd334292e672 | | ScanQA | 21.1 | Not recorded | 691ff208d6b0dd334292e672 | | FG-CLIP | 20.9 | Not recorded | 69d33e187d0172140ec7377f | | FG-CLIP | 20.9 | Not recorded | 691ff208d6b0dd334292e672 | | LLaVA-OV-7B | 20.8 | Not recorded | 691ff208d6b0dd334292e672 | ### Video-3D LLM EM@1: 30.1. Result ID: 69fc5f8c47f34ce924df83d3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modality":"RGB-D"}. This record does not assert independent reproduction. ### LLaVA-3D EM@1: 27. Result ID: 69fc5f8c47f34ce924df83d2. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modality":"RGB-D"}. This record does not assert independent reproduction. ### LEO EM@1: 24.5. Result ID: 69fc5f8c47f34ce924df83d1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modality":"C-PC"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B EM@1: 23.7. Result ID: 69fc5f8c47f34ce924df83d4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"RGB"}. This record does not assert independent reproduction. ### UniScene3D EM@1: 23.2. Result ID: 69d3651b47f34ce924baf751. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input":"CPM"}. This record does not assert independent reproduction. ### SceneVerse EM@1: 22.7. Result ID: 69d3651b47f34ce924baf74f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input":"PC"}. This record does not assert independent reproduction. ### SceneVerse EM@1: 22.7. Result ID: 69fc5f8c47f34ce924df83da. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"C-PC"}. This record does not assert independent reproduction. ### 3D-VisTA EM@1: 22.4. Result ID: 69d3651b47f34ce924baf74e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Input":"PC"}. This record does not assert independent reproduction. ### SplatTalk EM@1: 22.4. Result ID: 69fc5f8c47f34ce924df83d6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"RGB(GS)","Grid inputs":"H x W x 3","LoRA-tuned":"true"}. This record does not assert independent reproduction. ### 3D-ViSTA EM@1: 22.4. Result ID: 69fc5f8c47f34ce924df83d9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modality":"C-PC"}. This record does not assert independent reproduction. ### POMA-3D EM@1: 22.3. Result ID: 69d3651b47f34ce924baf750. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input":"PM"}. This record does not assert independent reproduction. ### POMA-3D\_spec EM@1: 22.3. Result ID: 69fc5f8c47f34ce924df83dc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"PM"}. This record does not assert independent reproduction. ### POMA-3D\_llm EM@1: 21.3. Result ID: 69fc5f8c47f34ce924df83d7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"PM","Grid inputs":"H x W x 3","LoRA-tuned":"true"}. This record does not assert independent reproduction. ### ScanQA EM@1: 21.1. Result ID: 69fc5f8c47f34ce924df83d8. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modality":"C-PC"}. This record does not assert independent reproduction. ### FG-CLIP EM@1: 20.9. Result ID: 69d3651b47f34ce924baf74d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input":"I"}. This record does not assert independent reproduction. ### FG-CLIP EM@1: 20.9. Result ID: 69fc5f8c47f34ce924df83db. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Modality":"PM"}. This record does not assert independent reproduction. ### LLaVA-OV-7B EM@1: 20.8. Result ID: 69fc5f8c47f34ce924df83d5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Modality":"RGB"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.