# Robot Video Diagnosis on RoboGazeBench (test) Source: [Robot Video Diagnosis on RoboGazeBench (test)]() Collection index: [llms.txt]() Data updated: 2026-06-30T06:22:41.060Z Evaluation of model performance on the test split of the RoboGazeBench dataset for Robot Video Diagnosis tasks, measured using metrics like Desc. F1, mIoU, F x IoU, and Clean 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 | Robot Video Diagnosis | | Dataset (including any stated split) | RoboGazeBench (test) | | Metric | Desc. F1 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 25 of 25 | | 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 - [Desc. F1](): Higher is better; unit: PERCENT. - [mIoU](): Higher is better; unit: PERCENT. - [F x IoU](): Higher is better; unit: PERCENT. - [Clean Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Desc. F1 | Result date | Report paper | | --- | --- | --- | --- | | Human ceiling | 75.5 | Not recorded | 6a434d70ba73165dbb7437fd | | Gemini 3.1 Pro | 67.9 | Not recorded | 6a434d70ba73165dbb7437fd | | GPT-5.5 | 65.1 | Not recorded | 6a434d70ba73165dbb7437fd | | Claude Sonnet 4.6 | 60.3 | Not recorded | 6a434d70ba73165dbb7437fd | | Gemini 3.1 Flash | 57.8 | Not recorded | 6a434d70ba73165dbb7437fd | | Gemma4-31B | 56.6 | Not recorded | 6a434d70ba73165dbb7437fd | | Qwen3.6-35B | 53.1 | Not recorded | 6a434d70ba73165dbb7437fd | | InternVL3.5-38B | 50.1 | Not recorded | 6a434d70ba73165dbb7437fd | | LLaVA-OV-2-8B | 47.9 | Not recorded | 6a434d70ba73165dbb7437fd | | Gemini 3.1 Pro | 32.4 | Not recorded | 6a434d70ba73165dbb7437fd | | GPT-5.5 | 30.5 | Not recorded | 6a434d70ba73165dbb7437fd | | Claude Sonnet 4.6 | 27.6 | Not recorded | 6a434d70ba73165dbb7437fd | | Gemini 3.1 Flash | 26.1 | Not recorded | 6a434d70ba73165dbb7437fd | | Gemini 3.1 Pro | 25.4 | Not recorded | 6a434d70ba73165dbb7437fd | | GPT-5.5 | 23.9 | Not recorded | 6a434d70ba73165dbb7437fd | | Gemma4-31B | 23.4 | Not recorded | 6a434d70ba73165dbb7437fd | | Qwen3.6-35B | 22.3 | Not recorded | 6a434d70ba73165dbb7437fd | | Claude Sonnet 4.6 | 21.4 | Not recorded | 6a434d70ba73165dbb7437fd | | InternVL3.5-38B | 21 | Not recorded | 6a434d70ba73165dbb7437fd | | Gemini 3.1 Flash | 20.2 | Not recorded | 6a434d70ba73165dbb7437fd | | LLaVA-OV-2-8B | 19.5 | Not recorded | 6a434d70ba73165dbb7437fd | | Gemma4-31B | 17.6 | Not recorded | 6a434d70ba73165dbb7437fd | | Qwen3.6-35B | 16.7 | Not recorded | 6a434d70ba73165dbb7437fd | | InternVL3.5-38B | 15.5 | Not recorded | 6a434d70ba73165dbb7437fd | | LLaVA-OV-2-8B | 14.2 | Not recorded | 6a434d70ba73165dbb7437fd | ### Human ceiling Desc. F1: 75.5. Result ID: 6a436095fbc373592c927d87. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Gemini 3.1 Pro Desc. F1: 67.9. Result ID: 6a436095fbc373592c927d71. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"ROBOGAZE (RG)"}. This record does not assert independent reproduction. ### GPT-5.5 Desc. F1: 65.1. Result ID: 6a436095fbc373592c927d74. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"ROBOGAZE (RG)"}. This record does not assert independent reproduction. ### Claude Sonnet 4.6 Desc. F1: 60.3. Result ID: 6a436095fbc373592c927d7a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"ROBOGAZE (RG)"}. This record does not assert independent reproduction. ### Gemini 3.1 Flash Desc. F1: 57.8. Result ID: 6a436095fbc373592c927d77. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"ROBOGAZE (RG)"}. This record does not assert independent reproduction. ### Gemma4-31B Desc. F1: 56.6. Result ID: 6a436095fbc373592c927d7d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"ROBOGAZE (RG)"}. This record does not assert independent reproduction. ### Qwen3.6-35B Desc. F1: 53.1. Result ID: 6a436095fbc373592c927d80. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"ROBOGAZE (RG)"}. This record does not assert independent reproduction. ### InternVL3.5-38B Desc. F1: 50.1. Result ID: 6a436095fbc373592c927d86. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"ROBOGAZE (RG)"}. This record does not assert independent reproduction. ### LLaVA-OV-2-8B Desc. F1: 47.9. Result ID: 6a436095fbc373592c927d83. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"ROBOGAZE (RG)"}. This record does not assert independent reproduction. ### Gemini 3.1 Pro Desc. F1: 32.4. Result ID: 6a436095fbc373592c927d70. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"chain-of-thought (CoT)"}. This record does not assert independent reproduction. ### GPT-5.5 Desc. F1: 30.5. Result ID: 6a436095fbc373592c927d73. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"chain-of-thought (CoT)"}. This record does not assert independent reproduction. ### Claude Sonnet 4.6 Desc. F1: 27.6. Result ID: 6a436095fbc373592c927d79. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"chain-of-thought (CoT)"}. This record does not assert independent reproduction. ### Gemini 3.1 Flash Desc. F1: 26.1. Result ID: 6a436095fbc373592c927d76. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"chain-of-thought (CoT)"}. This record does not assert independent reproduction. ### Gemini 3.1 Pro Desc. F1: 25.4. Result ID: 6a436095fbc373592c927d6f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"vanilla (V)"}. This record does not assert independent reproduction. ### GPT-5.5 Desc. F1: 23.9. Result ID: 6a436095fbc373592c927d72. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"vanilla (V)"}. This record does not assert independent reproduction. ### Gemma4-31B Desc. F1: 23.4. Result ID: 6a436095fbc373592c927d7c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"chain-of-thought (CoT)"}. This record does not assert independent reproduction. ### Qwen3.6-35B Desc. F1: 22.3. Result ID: 6a436095fbc373592c927d7f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"chain-of-thought (CoT)"}. This record does not assert independent reproduction. ### Claude Sonnet 4.6 Desc. F1: 21.4. Result ID: 6a436095fbc373592c927d78. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"vanilla (V)"}. This record does not assert independent reproduction. ### InternVL3.5-38B Desc. F1: 21. Result ID: 6a436095fbc373592c927d85. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"chain-of-thought (CoT)"}. This record does not assert independent reproduction. ### Gemini 3.1 Flash Desc. F1: 20.2. Result ID: 6a436095fbc373592c927d75. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Prop.","Prompting Strategy":"vanilla (V)"}. This record does not assert independent reproduction. ### LLaVA-OV-2-8B Desc. F1: 19.5. Result ID: 6a436095fbc373592c927d82. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"chain-of-thought (CoT)"}. This record does not assert independent reproduction. ### Gemma4-31B Desc. F1: 17.6. Result ID: 6a436095fbc373592c927d7b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"vanilla (V)"}. This record does not assert independent reproduction. ### Qwen3.6-35B Desc. F1: 16.7. Result ID: 6a436095fbc373592c927d7e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"vanilla (V)"}. This record does not assert independent reproduction. ### InternVL3.5-38B Desc. F1: 15.5. Result ID: 6a436095fbc373592c927d84. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"vanilla (V)"}. This record does not assert independent reproduction. ### LLaVA-OV-2-8B Desc. F1: 14.2. Result ID: 6a436095fbc373592c927d81. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Family":"Open","Prompting Strategy":"vanilla (V)"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.