# Lie Detection on MISID Source: [Lie Detection on MISID]() Collection index: [llms.txt]() Data updated: 2026-04-15T09:42:46.065Z Evaluation of Lie Detection performance on the MISID dataset, reporting Lie Detection Accuracy and Lie Reasoning 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 | Lie Detection | | Dataset (including any stated split) | MISID | | Metric | Lie Detection Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 29 of 29 | | 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 - [Lie Detection Accuracy](): Higher is better; unit: PERCENT. - [Lie Reasoning Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Lie Detection Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Grok-4-Fast | 61.54 | Not recorded | 69df1bbc7d0172140ec7badd | | Claude-Sonnet-4.5 | 61.14 | Not recorded | 69df1bbc7d0172140ec7badd | | GPT-5.1 | 60.36 | Not recorded | 69df1bbc7d0172140ec7badd | | Grok-4-Fast | 59.8 | Not recorded | 69df1bbc7d0172140ec7badd | | Claude-Sonnet-4.5 | 57.69 | Not recorded | 69df1bbc7d0172140ec7badd | | DeepSeek-V3 | 56.6 | Not recorded | 69df1bbc7d0172140ec7badd | | Gemini-3-flash | 53.88 | Not recorded | 69df1bbc7d0172140ec7badd | | GPT-5.1 | 53.85 | Not recorded | 69df1bbc7d0172140ec7badd | | GPT-4o | 53.85 | Not recorded | 69df1bbc7d0172140ec7badd | | Grok-4-Fast | 53.82 | Not recorded | 69df1bbc7d0172140ec7badd | | DeepSeek-V3 | 50.33 | Not recorded | 69df1bbc7d0172140ec7badd | | DeepSeek-R1 | 50.25 | Not recorded | 69df1bbc7d0172140ec7badd | | GPT-4o | 50.16 | Not recorded | 69df1bbc7d0172140ec7badd | | Gemini-2.5-flash | 50.16 | Not recorded | 69df1bbc7d0172140ec7badd | | GPT-4o | 49.84 | Not recorded | 69df1bbc7d0172140ec7badd | | Qwen3-VL-Flash | 46.15 | Not recorded | 69df1bbc7d0172140ec7badd | | Qwen3-VL-Max | 46.15 | Not recorded | 69df1bbc7d0172140ec7badd | | GLM-4.7 | 42.56 | Not recorded | 69df1bbc7d0172140ec7badd | | Qwen3-VL-235B-A22B | 42.31 | Not recorded | 69df1bbc7d0172140ec7badd | | DeepSeek-R1 | 42.31 | Not recorded | 69df1bbc7d0172140ec7badd | | Qwen3.5-Plus | 41.7 | Not recorded | 69df1bbc7d0172140ec7badd | | Qwen3-Max | 41.09 | Not recorded | 69df1bbc7d0172140ec7badd | | Qwen3-VL-Plus | 34.62 | Not recorded | 69df1bbc7d0172140ec7badd | | GLM-4.7 | 34.62 | Not recorded | 69df1bbc7d0172140ec7badd | | Qwen3-Max | 34.62 | Not recorded | 69df1bbc7d0172140ec7badd | | Qwen3.5-Plus | 34.62 | Not recorded | 69df1bbc7d0172140ec7badd | | MiniMax-M2.1 | 31.59 | Not recorded | 69df1bbc7d0172140ec7badd | | Qwen3.5-Plus | 30.77 | Not recorded | 69df1bbc7d0172140ec7badd | | MiniMax-M2.1 | 23.08 | Not recorded | 69df1bbc7d0172140ec7badd | ### Grok-4-Fast Lie Detection Accuracy: 61.54. Result ID: 69df5c8647f34ce924c87a60. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. ### Claude-Sonnet-4.5 Lie Detection Accuracy: 61.14. Result ID: 69df5c8647f34ce924c87a64. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### GPT-5.1 Lie Detection Accuracy: 60.36. Result ID: 69df5c8647f34ce924c87a69. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### Grok-4-Fast Lie Detection Accuracy: 59.8. Result ID: 69df5c8647f34ce924c87a6a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### Claude-Sonnet-4.5 Lie Detection Accuracy: 57.69. Result ID: 69df5c8647f34ce924c87a5a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. ### DeepSeek-V3 Lie Detection Accuracy: 56.6. Result ID: 69df5c8647f34ce924c87a66. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### Gemini-3-flash Lie Detection Accuracy: 53.88. Result ID: 69df5c8647f34ce924c87a59. [Reporting paper]() Recorded evaluation configuration: {"Input Modality":"Video + Text"}. This record does not assert independent reproduction. ### GPT-5.1 Lie Detection Accuracy: 53.85. Result ID: 69df5c8647f34ce924c87a5f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. ### GPT-4o Lie Detection Accuracy: 53.85. Result ID: 69df5c8647f34ce924c87a68. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### Grok-4-Fast Lie Detection Accuracy: 53.82. Result ID: 69df5c8647f34ce924c87a52. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Video + Text"}. This record does not assert independent reproduction. ### DeepSeek-V3 Lie Detection Accuracy: 50.33. Result ID: 69df5c8647f34ce924c87a5c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. ### DeepSeek-R1 Lie Detection Accuracy: 50.25. Result ID: 69df5c8647f34ce924c87a65. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### GPT-4o Lie Detection Accuracy: 50.16. Result ID: 69df5c8647f34ce924c87a51. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Video + Text"}. This record does not assert independent reproduction. ### Gemini-2.5-flash Lie Detection Accuracy: 50.16. Result ID: 69df5c8647f34ce924c87a58. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Video + Text"}. This record does not assert independent reproduction. ### GPT-4o Lie Detection Accuracy: 49.84. Result ID: 69df5c8647f34ce924c87a5e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. ### Qwen3-VL-Flash Lie Detection Accuracy: 46.15. Result ID: 69df5c8647f34ce924c87a54. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Video + Text"}. This record does not assert independent reproduction. ### Qwen3-VL-Max Lie Detection Accuracy: 46.15. Result ID: 69df5c8647f34ce924c87a55. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Video + Text"}. This record does not assert independent reproduction. ### GLM-4.7 Lie Detection Accuracy: 42.56. Result ID: 69df5c8647f34ce924c87a67. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### Qwen3-VL-235B-A22B Lie Detection Accuracy: 42.31. Result ID: 69df5c8647f34ce924c87a53. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Video + Text"}. This record does not assert independent reproduction. ### DeepSeek-R1 Lie Detection Accuracy: 42.31. Result ID: 69df5c8647f34ce924c87a5b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. ### Qwen3.5-Plus Lie Detection Accuracy: 41.7. Result ID: 69df5c8647f34ce924c87a6d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### Qwen3-Max Lie Detection Accuracy: 41.09. Result ID: 69df5c8647f34ce924c87a6c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### Qwen3-VL-Plus Lie Detection Accuracy: 34.62. Result ID: 69df5c8647f34ce924c87a56. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Video + Text"}. This record does not assert independent reproduction. ### GLM-4.7 Lie Detection Accuracy: 34.62. Result ID: 69df5c8647f34ce924c87a5d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. ### Qwen3-Max Lie Detection Accuracy: 34.62. Result ID: 69df5c8647f34ce924c87a62. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. ### Qwen3.5-Plus Lie Detection Accuracy: 34.62. Result ID: 69df5c8647f34ce924c87a63. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. ### MiniMax-M2.1 Lie Detection Accuracy: 31.59. Result ID: 69df5c8647f34ce924c87a6b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only","Assistance Framework":"FRACTAM"}. This record does not assert independent reproduction. ### Qwen3.5-Plus Lie Detection Accuracy: 30.77. Result ID: 69df5c8647f34ce924c87a57. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Video + Text"}. This record does not assert independent reproduction. ### MiniMax-M2.1 Lie Detection Accuracy: 23.08. Result ID: 69df5c8647f34ce924c87a61. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Input Modality":"Text-only"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.