# Anomaly Type Classification on M3-AD Average across scenes Source: [Anomaly Type Classification on M3-AD Average across scenes]() Collection index: [llms.txt]() Data updated: 2026-03-04T08:55:03.254Z Evaluation of Anomaly Type Classification performance across all scenes in the M3-AD dataset, reporting the Type metric. 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 Type Classification | | Dataset (including any stated split) | M3-AD Average across scenes | | Metric | Type Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 19 of 19 | | 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 - [Type Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Type Score | Result date | Report paper | | --- | --- | --- | --- | | QWEN-3-VL-INSTRUCT | 52.7 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-3-VL-INSTRUCT | 51.2 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | GPT-5.1-MINI | 49.8 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | GEMINI-2.5-FLASH-LITE | 44.4 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | GPT-5.1-NANO | 37.4 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-3-VL-INSTRUCT | 36.8 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | INTERN-VL-3.5 | 36.2 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-3-VL-THINKING | 34.9 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-3-VL-THINKING | 34.2 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | INTERN-VL-3.5 | 32.7 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-3-VL-INSTRUCT | 32.6 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-3-VL-INSTRUCT | 27.1 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-2.5-VL-INSTRUCT | 25.5 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | LLAVA-ONEVISION-SI | 24.2 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-2.5-VL-INSTRUCT | 22.5 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | ANOMALY-R1 | 20.6 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-2.5-VL-INSTRUCT | 20.3 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-2-VL-INSTRUCT | 16.9 | Not recorded | 69a66bdb2d603e6de3e3cc6e | | QWEN-2-VL-INSTRUCT | 1.2 | Not recorded | 69a66bdb2d603e6de3e3cc6e | ### QWEN-3-VL-INSTRUCT Type Score: 52.7. Result ID: 69a7f35c6dbb7075b94ceaec. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"RA-MONITOR","Param":"4B"}. This record does not assert independent reproduction. ### QWEN-3-VL-INSTRUCT Type Score: 51.2. Result ID: 69a7f35c6dbb7075b94ceaed. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"RA-MONITOR","Param":"8B"}. This record does not assert independent reproduction. ### GPT-5.1-MINI Type Score: 49.8. Result ID: 69a7f35c6dbb7075b94ceadc. [Reporting paper]() Recorded evaluation configuration: {"Category":"COMMERCIAL","Param":"/"}. This record does not assert independent reproduction. ### GEMINI-2.5-FLASH-LITE Type Score: 44.4. Result ID: 69a7f35c6dbb7075b94ceadd. [Reporting paper]() Recorded evaluation configuration: {"Category":"COMMERCIAL","Param":"/"}. This record does not assert independent reproduction. ### GPT-5.1-NANO Type Score: 37.4. Result ID: 69a7f35c6dbb7075b94ceadb. [Reporting paper]() Recorded evaluation configuration: {"Category":"COMMERCIAL","Param":"/"}. This record does not assert independent reproduction. ### QWEN-3-VL-INSTRUCT Type Score: 36.8. Result ID: 69a7f35c6dbb7075b94ceae8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"32B"}. This record does not assert independent reproduction. ### INTERN-VL-3.5 Type Score: 36.2. Result ID: 69a7f35c6dbb7075b94ceae4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"4B"}. This record does not assert independent reproduction. ### QWEN-3-VL-THINKING Type Score: 34.9. Result ID: 69a7f35c6dbb7075b94ceaea. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"THINKING","Param":"4B"}. This record does not assert independent reproduction. ### QWEN-3-VL-THINKING Type Score: 34.2. Result ID: 69a7f35c6dbb7075b94ceaeb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"THINKING","Param":"8B"}. This record does not assert independent reproduction. ### INTERN-VL-3.5 Type Score: 32.7. Result ID: 69a7f35c6dbb7075b94ceae5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"8B"}. This record does not assert independent reproduction. ### QWEN-3-VL-INSTRUCT Type Score: 32.6. Result ID: 69a7f35c6dbb7075b94ceae6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"4B"}. This record does not assert independent reproduction. ### QWEN-3-VL-INSTRUCT Type Score: 27.1. Result ID: 69a7f35c6dbb7075b94ceae7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"8B"}. This record does not assert independent reproduction. ### QWEN-2.5-VL-INSTRUCT Type Score: 25.5. Result ID: 69a7f35c6dbb7075b94ceae9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"72B"}. This record does not assert independent reproduction. ### LLAVA-ONEVISION-SI Type Score: 24.2. Result ID: 69a7f35c6dbb7075b94ceae3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"7B"}. This record does not assert independent reproduction. ### QWEN-2.5-VL-INSTRUCT Type Score: 22.5. Result ID: 69a7f35c6dbb7075b94ceae2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"7B"}. This record does not assert independent reproduction. ### ANOMALY-R1 Type Score: 20.6. Result ID: 69a7f35c6dbb7075b94ceade. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"3B"}. This record does not assert independent reproduction. ### QWEN-2.5-VL-INSTRUCT Type Score: 20.3. Result ID: 69a7f35c6dbb7075b94ceae1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"3B"}. This record does not assert independent reproduction. ### QWEN-2-VL-INSTRUCT Type Score: 16.9. Result ID: 69a7f35c6dbb7075b94ceae0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"7B"}. This record does not assert independent reproduction. ### QWEN-2-VL-INSTRUCT Type Score: 1.2. Result ID: 69a7f35c6dbb7075b94ceadf. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Category":"OPEN-SOURCE","Param":"2B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.