# Tabular Question Answering on ReasonTabQA 1.0 (Overall) Source: [Tabular Question Answering on ReasonTabQA 1.0 (Overall)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:54:25.540Z Evaluation of model performance on the ReasonTabQA 1.0 dataset (Overall split), reporting accuracy broken down by language (English/Chinese) and difficulty levels (Question and Table difficulty). 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 | Tabular Question Answering | | Dataset (including any stated split) | ReasonTabQA 1.0 (Overall) | | Metric | Overall Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 33 of 33 | | 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 Accuracy](): Higher is better; unit: PERCENT. - [Accuracy (English)](): Higher is better; unit: PERCENT. - [Accuracy (Chinese)](): Higher is better; unit: PERCENT. - [Accuracy (Q-Easy)](): Higher is better; unit: PERCENT. - [Accuracy (Q-Medium)](): Higher is better; unit: PERCENT. - [Accuracy (Q-Hard)](): Higher is better; unit: PERCENT. - [Accuracy (T-Simple)](): Higher is better; unit: PERCENT. - [Accuracy (T-Medium)](): Higher is better; unit: PERCENT. - [Accuracy (T-Complex)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Overall Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Gemini-3-Pro-Preview | 67.58 | Not recorded | 6965d19a6343fa78c832367d | | Claude-Opus-4.5 | 66.2 | Not recorded | 6965d19a6343fa78c832367d | | Doubao-1.5-thinking-pro | 64.48 | Not recorded | 6965d19a6343fa78c832367d | | Claude-4.0-Sonnet | 62.53 | Not recorded | 6965d19a6343fa78c832367d | | Qwen3-8B-Think-SFT-TabCodeRL | 61.89 | Not recorded | 6965d19a6343fa78c832367d | | Kimi-K2-Instruct | 59.57 | Not recorded | 6965d19a6343fa78c832367d | | GPT-5.2 | 59.46 | Not recorded | 6965d19a6343fa78c832367d | | Qwen3-32B (thinking) | 58.76 | Not recorded | 6965d19a6343fa78c832367d | | Qwen3-8B-NoThink-SFT-TabCodeRL | 58.01 | Not recorded | 6965d19a6343fa78c832367d | | GPT-4o | 56.9 | Not recorded | 6965d19a6343fa78c832367d | | Deepseek-R1 | 55.8 | Not recorded | 6965d19a6343fa78c832367d | | OpenAI o1-mini | 55.8 | Not recorded | 6965d19a6343fa78c832367d | | QWQ-32B | 54.1 | Not recorded | 6965d19a6343fa78c832367d | | DeepSeek-R1-Dist-Qwen-32B | 53.64 | Not recorded | 6965d19a6343fa78c832367d | | Qwen3-30B-A3B (thinking) | 53.61 | Not recorded | 6965d19a6343fa78c832367d | | Qwen3-32B (no-thinking) | 53.13 | Not recorded | 6965d19a6343fa78c832367d | | Deepseek-V3 | 52.15 | Not recorded | 6965d19a6343fa78c832367d | | Qwen2.5-72B-Instruct | 51.92 | Not recorded | 6965d19a6343fa78c832367d | | Telechat3-36B | 51.13 | Not recorded | 6965d19a6343fa78c832367d | | Qwen3-8B (thinking) | 49.87 | Not recorded | 6965d19a6343fa78c832367d | | Mistral-Large-Instruct-2407 | 47.17 | Not recorded | 6965d19a6343fa78c832367d | | Qwen2.5-Coder-32B-Instruct | 45.28 | Not recorded | 6965d19a6343fa78c832367d | | TableGPT2-7B | 42.05 | Not recorded | 6965d19a6343fa78c832367d | | Qwen3-8B (no-thinking) | 40.97 | Not recorded | 6965d19a6343fa78c832367d | | DeepSeek-R1-Dist-Qwen-14B | 39.89 | Not recorded | 6965d19a6343fa78c832367d | | Llama-3.3-70B-Instruct | 37.74 | Not recorded | 6965d19a6343fa78c832367d | | Qwen2-72B-Instruct | 36.66 | Not recorded | 6965d19a6343fa78c832367d | | DS-R1-Dist-Qwen-7B-SFT-TabCodeRL | 34.46 | Not recorded | 6965d19a6343fa78c832367d | | Llama-3.1-8B-Instruct | 33.24 | Not recorded | 6965d19a6343fa78c832367d | | TableLlama-7B-SFT-TabCodeRL | 20.43 | Not recorded | 6965d19a6343fa78c832367d | | TableLLM-7B | 13.5 | Not recorded | 6965d19a6343fa78c832367d | | TableLlama-7B | 13.23 | Not recorded | 6965d19a6343fa78c832367d | | DeepSeek-R1-Dist-Qwen-7B | 11.86 | Not recorded | 6965d19a6343fa78c832367d | ### Gemini-3-Pro-Preview Overall Accuracy: 67.58. Result ID: 69968e825698665ec4d759a9. [Reporting paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Closed-Source"}. This record does not assert independent reproduction. ### Claude-Opus-4.5 Overall Accuracy: 66.2. Result ID: 69968e825698665ec4d759a8. [Reporting paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Closed-Source"}. This record does not assert independent reproduction. ### Doubao-1.5-thinking-pro Overall Accuracy: 64.48. Result ID: 69968e825698665ec4d759a5. [Reporting paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Closed-Source"}. This record does not assert independent reproduction. ### Claude-4.0-Sonnet Overall Accuracy: 62.53. Result ID: 69968e825698665ec4d759a7. [Reporting paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Closed-Source"}. This record does not assert independent reproduction. ### Qwen3-8B-Think-SFT-TabCodeRL Overall Accuracy: 61.89. Result ID: 69968e825698665ec4d759ad. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Proposed Framework":"TabCodeRL","Base Model":"Qwen3-8B-Think"}. This record does not assert independent reproduction. ### Kimi-K2-Instruct Overall Accuracy: 59.57. Result ID: 69968e825698665ec4d75997. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### GPT-5.2 Overall Accuracy: 59.46. Result ID: 69968e825698665ec4d75999. [Reporting paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Closed-Source"}. This record does not assert independent reproduction. ### Qwen3-32B (thinking) Overall Accuracy: 58.76. Result ID: 69968e825698665ec4d7599f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Qwen3-8B-NoThink-SFT-TabCodeRL Overall Accuracy: 58.01. Result ID: 69968e825698665ec4d759ac. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Proposed Framework":"TabCodeRL","Base Model":"Qwen3-8B-NoThink"}. This record does not assert independent reproduction. ### GPT-4o Overall Accuracy: 56.9. Result ID: 69968e825698665ec4d75998. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Closed-Source"}. This record does not assert independent reproduction. ### Deepseek-R1 Overall Accuracy: 55.8. Result ID: 69968e825698665ec4d759a4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### OpenAI o1-mini Overall Accuracy: 55.8. Result ID: 69968e825698665ec4d759a6. [Reporting paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Closed-Source"}. This record does not assert independent reproduction. ### QWQ-32B Overall Accuracy: 54.1. Result ID: 69968e825698665ec4d7599d. [Reporting paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### DeepSeek-R1-Dist-Qwen-32B Overall Accuracy: 53.64. Result ID: 69968e825698665ec4d759a3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Qwen3-30B-A3B (thinking) Overall Accuracy: 53.61. Result ID: 69968e825698665ec4d759a0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Qwen3-32B (no-thinking) Overall Accuracy: 53.13. Result ID: 69968e825698665ec4d75991. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Deepseek-V3 Overall Accuracy: 52.15. Result ID: 69968e825698665ec4d75996. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Qwen2.5-72B-Instruct Overall Accuracy: 51.92. Result ID: 69968e825698665ec4d7598f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Telechat3-36B Overall Accuracy: 51.13. Result ID: 69968e825698665ec4d75995. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Qwen3-8B (thinking) Overall Accuracy: 49.87. Result ID: 69968e825698665ec4d7599e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Mistral-Large-Instruct-2407 Overall Accuracy: 47.17. Result ID: 69968e825698665ec4d75994. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Qwen2.5-Coder-32B-Instruct Overall Accuracy: 45.28. Result ID: 69968e825698665ec4d7598e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### TableGPT2-7B Overall Accuracy: 42.05. Result ID: 69968e825698665ec4d7599a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Specialization":"Table-Specific"}. This record does not assert independent reproduction. ### Qwen3-8B (no-thinking) Overall Accuracy: 40.97. Result ID: 69968e825698665ec4d75990. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### DeepSeek-R1-Dist-Qwen-14B Overall Accuracy: 39.89. Result ID: 69968e825698665ec4d759a2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Llama-3.3-70B-Instruct Overall Accuracy: 37.74. Result ID: 69968e825698665ec4d75993. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### Qwen2-72B-Instruct Overall Accuracy: 36.66. Result ID: 69968e825698665ec4d7598d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### DS-R1-Dist-Qwen-7B-SFT-TabCodeRL Overall Accuracy: 34.46. Result ID: 69968e825698665ec4d759ab. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Proposed Framework":"TabCodeRL","Base Model":"DeepSeek-R1-Dist-Qwen-7B"}. This record does not assert independent reproduction. ### Llama-3.1-8B-Instruct Overall Accuracy: 33.24. Result ID: 69968e825698665ec4d75992. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. ### TableLlama-7B-SFT-TabCodeRL Overall Accuracy: 20.43. Result ID: 69968e825698665ec4d759aa. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Proposed Framework":"TabCodeRL","Base Model":"TableLlama-7B"}. This record does not assert independent reproduction. ### TableLLM-7B Overall Accuracy: 13.5. Result ID: 69968e825698665ec4d7599b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Specialization":"Table-Specific"}. This record does not assert independent reproduction. ### TableLlama-7B Overall Accuracy: 13.23. Result ID: 69968e825698665ec4d7599c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"No Reasoning","Model Specialization":"Table-Specific"}. This record does not assert independent reproduction. ### DeepSeek-R1-Dist-Qwen-7B Overall Accuracy: 11.86. Result ID: 69968e825698665ec4d759a1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Reasoning Mechanism":"Reasoning","Model Source":"Open-Source"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.