# Anomaly Discrimination on Anomaly Discrimination Tasks Source: [Anomaly Discrimination on Anomaly Discrimination Tasks]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:31:21.828Z Evaluation of Anomaly Discrimination performance on the Anomaly Discrimination Tasks dataset, reporting Accuracy, Recall, Precision, and F1 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 | Anomaly Discrimination | | Dataset (including any stated split) | Anomaly Discrimination Tasks | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 28 of 28 | | 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. - [Recall](): Higher is better; unit: PERCENT. - [Precision](): Higher is better; unit: PERCENT. - [F1 Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Human (expert) | 95.24 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Human (ordinary) | 86.9 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Reason-IAD (Qwen3-VL) | 79.42 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Reason-IAD (Qwen3-VL) | 79.2 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Qwen3-VL | 74.51 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Qwen3-VL | 74.48 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Reason-IAD (Qwen2.5-VL) | 73.91 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Reason-IAD (Qwen3-VL) | 73.48 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Qwen3-VL | 69.24 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Qwen2.5-VL | 68.87 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Gemini-1.5-pro | 68.63 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | GPT-4o | 68.63 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | InternVL2 | 68.25 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | AnomalyGPT | 65.57 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | GPT-4o-mini | 64.33 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | claude-3.5-sonnet | 60.14 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | InternVL2 | 59.97 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Gemini-1.5-flash | 58.58 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | LLaVA-NeXT | 57.92 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | LLaVA-NeXT-Interleave | 57.64 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | MiniCPM-V2.6 | 57.31 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | InternLM-XComposer2-VL | 55.85 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Cambrian-1\* | 55.6 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | Qwen-VL-Chat | 53.65 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | SPHINX\* | 53.13 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | LLaVA-OnVision | 51.77 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | LLaVA-1.5 | 51.33 | Not recorded | 698c0d22c97a0b3e3b5e11cf | | LLaVA-1.5 | 49.96 | Not recorded | 698c0d22c97a0b3e3b5e11cf | ### Human (expert) Accuracy: 95.24. Result ID: 69971bb55698665ec469b3e8. [Reporting paper]() Recorded evaluation configuration: {"Scale":"-"}. This record does not assert independent reproduction. ### Human (ordinary) Accuracy: 86.9. Result ID: 69971bb55698665ec469b3e9. [Reporting paper]() Recorded evaluation configuration: {"Scale":"-"}. This record does not assert independent reproduction. ### Reason-IAD (Qwen3-VL) Accuracy: 79.42. Result ID: 69971bb55698665ec469b403. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"8B"}. This record does not assert independent reproduction. ### Reason-IAD (Qwen3-VL) Accuracy: 79.2. Result ID: 69971bb55698665ec469b401. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"4B"}. This record does not assert independent reproduction. ### Qwen3-VL Accuracy: 74.51. Result ID: 69971bb55698665ec469b3f5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"8B"}. This record does not assert independent reproduction. ### Qwen3-VL Accuracy: 74.48. Result ID: 69971bb55698665ec469b3f0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"4B"}. This record does not assert independent reproduction. ### Reason-IAD (Qwen2.5-VL) Accuracy: 73.91. Result ID: 69971bb55698665ec469b402. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### Reason-IAD (Qwen3-VL) Accuracy: 73.48. Result ID: 69971bb55698665ec469b400. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"2B"}. This record does not assert independent reproduction. ### Qwen3-VL Accuracy: 69.24. Result ID: 69971bb55698665ec469b3ef. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"2B"}. This record does not assert independent reproduction. ### Qwen2.5-VL Accuracy: 68.87. Result ID: 69971bb55698665ec469b3f2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### Gemini-1.5-pro Accuracy: 68.63. Result ID: 69971bb55698665ec469b3ec. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"-"}. This record does not assert independent reproduction. ### GPT-4o Accuracy: 68.63. Result ID: 69971bb55698665ec469b3ee. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"-"}. This record does not assert independent reproduction. ### InternVL2 Accuracy: 68.25. Result ID: 69971bb55698665ec469b3ff. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Scale":"76B"}. This record does not assert independent reproduction. ### AnomalyGPT Accuracy: 65.57. Result ID: 69971bb55698665ec469b3f1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### GPT-4o-mini Accuracy: 64.33. Result ID: 69971bb55698665ec469b3ed. [Reporting paper]() Recorded evaluation configuration: {"Scale":"-"}. This record does not assert independent reproduction. ### claude-3.5-sonnet Accuracy: 60.14. Result ID: 69971bb55698665ec469b3ea. [Reporting paper]() Recorded evaluation configuration: {"Scale":"-"}. This record does not assert independent reproduction. ### InternVL2 Accuracy: 59.97. Result ID: 69971bb55698665ec469b3fc. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Scale":"8B"}. This record does not assert independent reproduction. ### Gemini-1.5-flash Accuracy: 58.58. Result ID: 69971bb55698665ec469b3eb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Scale":"-"}. This record does not assert independent reproduction. ### LLaVA-NeXT Accuracy: 57.92. Result ID: 69971bb55698665ec469b3fe. [Reporting paper]() Recorded evaluation configuration: {"Scale":"34B"}. This record does not assert independent reproduction. ### LLaVA-NeXT-Interleave Accuracy: 57.64. Result ID: 69971bb55698665ec469b3f8. [Reporting paper]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### MiniCPM-V2.6 Accuracy: 57.31. Result ID: 69971bb55698665ec469b3fb. [Reporting paper]() Recorded evaluation configuration: {"Scale":"8B"}. This record does not assert independent reproduction. ### InternLM-XComposer2-VL Accuracy: 55.85. Result ID: 69971bb55698665ec469b3f9. [Reporting paper]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### Cambrian-1\* Accuracy: 55.6. Result ID: 69971bb55698665ec469b3f6. [Reporting paper]() Recorded evaluation configuration: {"Scale":"8B"}. This record does not assert independent reproduction. ### Qwen-VL-Chat Accuracy: 53.65. Result ID: 69971bb55698665ec469b3f3. [Reporting paper]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### SPHINX\* Accuracy: 53.13. Result ID: 69971bb55698665ec469b3f7. [Reporting paper]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### LLaVA-OnVision Accuracy: 51.77. Result ID: 69971bb55698665ec469b3fa. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### LLaVA-1.5 Accuracy: 51.33. Result ID: 69971bb55698665ec469b3f4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Scale":"7B"}. This record does not assert independent reproduction. ### LLaVA-1.5 Accuracy: 49.96. Result ID: 69971bb55698665ec469b3fd. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Scale":"13B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.