# multi-view Visual Question Answering on VSI-Bench (test) Source: [multi-view Visual Question Answering on VSI-Bench (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:35:51.451Z Evaluation of model performance on the test split of the VSI-Bench dataset for multi-view Visual Question Answering, reporting metrics related to spatial relationships, object counts, and route planning. 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 | multi-view Visual Question Answering | | Dataset (including any stated split) | VSI-Bench (test) | | Metric | Average Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 11 of 11 | | 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 - [Average Score](): Higher is better; unit: PERCENT. - [Object Count Score](): Higher is better; unit: SCALAR. - [Absolute Distance Error](): Lower is better; unit: OTHER. - [Object Size Score](): Higher is better; unit: SCALAR. - [Room Size Score](): Higher is better; unit: SCALAR. - [Relative Distance Error](): Lower is better; unit: OTHER. - [Relative Direction Error](): Lower is better; unit: OTHER. - [Route Planning Success Rate](): Higher is better; unit: PERCENT. - [Appearance Order Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Average Score | Result date | Report paper | | --- | --- | --- | --- | | VaLR-M | 52.9 | Not recorded | 6984241fa8b01e76cc8ee79d | | VaLR-S | 41.5 | Not recorded | 6984241fa8b01e76cc8ee79d | | LLaVA-NeXT-Video-7B | 35.6 | Not recorded | 6984241fa8b01e76cc8ee79d | | GPT-4o | 34 | Not recorded | 6984241fa8b01e76cc8ee79d | | Qwen2.5-VL-7B + vanilla SFT | 33.7 | Not recorded | 6984241fa8b01e76cc8ee79d | | Qwen2.5-VL-7B | 33 | Not recorded | 6984241fa8b01e76cc8ee79d | | Ocean-R1-7B | 30.5 | Not recorded | 6984241fa8b01e76cc8ee79d | | Qwen2.5-VL-7B + CoVT | 18.6 | Not recorded | 6984241fa8b01e76cc8ee79d | | Qwen2.5-VL-7B + LVR | 18.4 | Not recorded | 6984241fa8b01e76cc8ee79d | | R1-OneVision-7B | 16.1 | Not recorded | 6984241fa8b01e76cc8ee79d | | Qwen2.5-VL-7B + Monet | 14 | Not recorded | 6984241fa8b01e76cc8ee79d | ### VaLR-M Average Score: 52.9. Result ID: 699744cf5698665ec495cb91. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Group":"Latent Reasoning Models","Encoder Alignment":"Multiple encoder (DINOv3, SigLIPv2, pi3)"}. This record does not assert independent reproduction. ### VaLR-S Average Score: 41.5. Result ID: 699744cf5698665ec495cb90. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Group":"Latent Reasoning Models","Encoder Alignment":"Single encoder (DINOv3)"}. This record does not assert independent reproduction. ### LLaVA-NeXT-Video-7B Average Score: 35.6. Result ID: 699744cf5698665ec495cb88. [Reporting paper]() Recorded evaluation configuration: {"Method Group":"Other Models","Model Architecture/Scale":"7B"}. This record does not assert independent reproduction. ### GPT-4o Average Score: 34. Result ID: 699744cf5698665ec495cb87. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Group":"Other Models"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B + vanilla SFT Average Score: 33.7. Result ID: 699744cf5698665ec495cb8c. [Reporting paper]() Recorded evaluation configuration: {"Method Group":"Base Model","Training Strategy":"vanilla SFT"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B Average Score: 33. Result ID: 699744cf5698665ec495cb8b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Group":"Base Model","Model Architecture/Scale":"7B"}. This record does not assert independent reproduction. ### Ocean-R1-7B Average Score: 30.5. Result ID: 699744cf5698665ec495cb8a. [Reporting paper]() Recorded evaluation configuration: {"Method Group":"Reasoning Models","Model Architecture/Scale":"7B"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B + CoVT Average Score: 18.6. Result ID: 699744cf5698665ec495cb8e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Group":"Latent Reasoning Models"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B + LVR Average Score: 18.4. Result ID: 699744cf5698665ec495cb8d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Group":"Latent Reasoning Models"}. This record does not assert independent reproduction. ### R1-OneVision-7B Average Score: 16.1. Result ID: 699744cf5698665ec495cb89. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Group":"Reasoning Models","Model Architecture/Scale":"7B"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B + Monet Average Score: 14. Result ID: 699744cf5698665ec495cb8f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Method Group":"Latent Reasoning Models"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.