# Spatial Reasoning on CV-Bench (Component Metrics) Source: [Spatial Reasoning on CV-Bench (Component Metrics)]() Collection index: [llms.txt]() Data updated: 2026-05-29T11:34:34.578Z Evaluation of model performance on the CV-Bench dataset for Spatial Reasoning, reporting component scores for Count, Relation, Depth, and Distance, along with an Overall score. 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 | Spatial Reasoning | | Dataset (including any stated split) | CV-Bench | | Metric | Overall Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | 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 - [Overall Score](): Higher is better; unit: PERCENT. - [Count](): Higher is better; unit: PERCENT. - [Relation](): Higher is better; unit: PERCENT. - [Depth](): Higher is better; unit: PERCENT. - [Distance](): Higher is better; unit: PERCENT. - [Count (2D)](): Higher is better; unit: PERCENT. - [Relation (2D)](): Higher is better; unit: PERCENT. - [Depth (3D)](): Higher is better; unit: PERCENT. - [Distance (3D)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Overall Score | Result date | Report paper | | --- | --- | --- | --- | | SpatialStack | 86.5 | Not recorded | 69cb55020636e5bf4afc9ec4 | | GASP - Full Model (Lcorr + Ldepth) | 79.8 | Not recorded | 6a191dbcdbd7150506f829de | | Qwen2.5-VL-32B | 79.7 | Not recorded | 6a191dbcdbd7150506f829de | | LLaVA-OneVision-72B | 79.7 | Not recorded | 6a191dbcdbd7150506f829de | | GASP - Correspondence (Lcorr) | 79.4 | Not recorded | 6a191dbcdbd7150506f829de | | Qwen2.5-VL-7B-Instruct | 76.6 | Not recorded | 6a191dbcdbd7150506f829de | | InternVL2.5-4B | 74.1 | Not recorded | 6a191dbcdbd7150506f829de | | LLaVA-OneVision-7B | 73.2 | Not recorded | 6a191dbcdbd7150506f829de | | LLaMA-3.2V-11B-CoT | 72.8 | Not recorded | 6a191dbcdbd7150506f829de | | SpatialRGPT | 72.7 | Not recorded | 69cb55020636e5bf4afc9ec4 | | Qwen2.5-VL-3B-Instruct | 68.5 | Not recorded | 6a191dbcdbd7150506f829de | | Spatialbot | 68 | Not recorded | 69cb55020636e5bf4afc9ec4 | | Cambrain-8B | 62.2 | Not recorded | 6a191dbcdbd7150506f829de | | LLaMA-3.2V-11B | 58.2 | Not recorded | 6a191dbcdbd7150506f829de | | LLaVA-1.5-13B | 58.2 | Not recorded | 6a191dbcdbd7150506f829de | | SpaceLLaVA-13B | 58.2 | Not recorded | 6a191dbcdbd7150506f829de | | Kimi-VL-3B-Thinking | 57.5 | Not recorded | 6a191dbcdbd7150506f829de | | SpaceQwen2.5VL-3B | 51.4 | Not recorded | 6a191dbcdbd7150506f829de | ### SpatialStack Overall Score: 86.5. Result ID: 69cb8aa947f34ce924b3efdc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"zero-shot":"true"}. This record does not assert independent reproduction. ### GASP - Full Model (Lcorr + Ldepth) Overall Score: 79.8. Result ID: 6a1979c1fbc373592c14c9e6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction. ### Qwen2.5-VL-32B Overall Score: 79.7. Result ID: 6a1979c1fbc373592c14c9e3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### LLaVA-OneVision-72B Overall Score: 79.7. Result ID: 6a1979c1fbc373592c14c9e4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### GASP - Correspondence (Lcorr) Overall Score: 79.4. Result ID: 6a1979c1fbc373592c14c9e5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B-Instruct Overall Score: 76.6. Result ID: 6a1979c1fbc373592c14c9dd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### InternVL2.5-4B Overall Score: 74.1. Result ID: 6a1979c1fbc373592c14c9db. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### LLaVA-OneVision-7B Overall Score: 73.2. Result ID: 6a1979c1fbc373592c14c9dc. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### LLaMA-3.2V-11B-CoT Overall Score: 72.8. Result ID: 6a1979c1fbc373592c14c9e0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Chain-of-Thought":"true"}. This record does not assert independent reproduction. ### SpatialRGPT Overall Score: 72.7. Result ID: 69cb8aa947f34ce924b3efda. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"zero-shot":"true"}. This record does not assert independent reproduction. ### Qwen2.5-VL-3B-Instruct Overall Score: 68.5. Result ID: 6a1979c1fbc373592c14c9da. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Spatialbot Overall Score: 68. Result ID: 69cb8aa947f34ce924b3efdb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"zero-shot":"true"}. This record does not assert independent reproduction. ### Cambrain-8B Overall Score: 62.2. Result ID: 6a1979c1fbc373592c14c9de. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### LLaMA-3.2V-11B Overall Score: 58.2. Result ID: 6a1979c1fbc373592c14c9df. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### LLaVA-1.5-13B Overall Score: 58.2. Result ID: 6a1979c1fbc373592c14c9e1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### SpaceLLaVA-13B Overall Score: 58.2. Result ID: 6a1979c1fbc373592c14c9e2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Kimi-VL-3B-Thinking Overall Score: 57.5. Result ID: 6a1979c1fbc373592c14c9d9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### SpaceQwen2.5VL-3B Overall Score: 51.4. Result ID: 6a1979c1fbc373592c14c9d8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.