# Document Image Question Answering on InfographicVQA (test) Source: [Document Image Question Answering on InfographicVQA (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:22:33.698Z Evaluation of model performance on the test split of the InfographicVQA dataset specifically for Document Image Question Answering, reporting detailed span and category metrics alongside ANLS. 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 | Document Image Question Answering | | Dataset (including any stated split) | InfographicVQA (test) | | Metric | ANLS | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 10 of 10 | | 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 - [ANLS](): Higher is better; unit: SCALAR. - [Delta ANLS](): Higher is better; unit: SCALAR. - [Image Span Accuracy](): Higher is better; unit: PERCENT. - [Question Span Accuracy](): Higher is better; unit: PERCENT. - [Multiple Spans Accuracy](): Higher is better; unit: PERCENT. - [Non-Span Accuracy](): Higher is better; unit: PERCENT. - [Table/List Accuracy](): Higher is better; unit: PERCENT. - [Textual Accuracy](): Higher is better; unit: PERCENT. - [Visual Object Accuracy](): Higher is better; unit: PERCENT. - [Figure Accuracy](): Higher is better; unit: PERCENT. - [Map Accuracy](): Higher is better; unit: PERCENT. - [Comparison Accuracy](): Higher is better; unit: PERCENT. - [Arithmetic Accuracy](): Higher is better; unit: PERCENT. - [Counting Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | ANLS | Result date | Report paper | | --- | --- | --- | --- | | TILT | 0.612 | Not recorded | 68fa707c1e209131cea4b65b | | Claude + LATIN-Prompt | 0.5451 | Not recorded | 68fa707c1e209131cea4b65b | | ChatGPT-3.5 + LATIN-Prompt | 0.4898 | Not recorded | 68fa707c1e209131cea4b65b | | pix2struct | 0.4001 | Not recorded | 68fa707c1e209131cea4b65b | | ChatGPT-3.5 + Plain Prompt | 0.3335 | Not recorded | 68fa707c1e209131cea4b65b | | BROS | 0.3219 | Not recorded | 68fa707c1e209131cea4b65b | | LayoutLMv2 | 0.2829 | Not recorded | 68fa707c1e209131cea4b65b | | LayoutLM | 0.272 | Not recorded | 68fa707c1e209131cea4b65b | | BERT | 0.2078 | Not recorded | 68fa707c1e209131cea4b65b | | Claude + Plain Prompt | 0.0798 | Not recorded | 68fa707c1e209131cea4b65b | ### TILT ANLS: 0.612. Result ID: 6991445a3a6c31484906e882. [Reporting paper]() Recorded evaluation configuration: {"Paradigm":"Fine-tuning"}. This record does not assert independent reproduction. ### Claude + LATIN-Prompt ANLS: 0.5451. Result ID: 6991445a3a6c31484906e884. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Zero-shot"}. This record does not assert independent reproduction. ### ChatGPT-3.5 + LATIN-Prompt ANLS: 0.4898. Result ID: 6991445a3a6c31484906e886. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Zero-shot"}. This record does not assert independent reproduction. ### pix2struct ANLS: 0.4001. Result ID: 6991445a3a6c31484906e881. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Fine-tuning"}. This record does not assert independent reproduction. ### ChatGPT-3.5 + Plain Prompt ANLS: 0.3335. Result ID: 6991445a3a6c31484906e885. [Reporting paper]() Recorded evaluation configuration: {"Paradigm":"Zero-shot"}. This record does not assert independent reproduction. ### BROS ANLS: 0.3219. Result ID: 6991445a3a6c31484906e880. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Fine-tuning"}. This record does not assert independent reproduction. ### LayoutLMv2 ANLS: 0.2829. Result ID: 6991445a3a6c31484906e87f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Fine-tuning"}. This record does not assert independent reproduction. ### LayoutLM ANLS: 0.272. Result ID: 6991445a3a6c31484906e87e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Fine-tuning"}. This record does not assert independent reproduction. ### BERT ANLS: 0.2078. Result ID: 6991445a3a6c31484906e87d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Paradigm":"Fine-tuning"}. This record does not assert independent reproduction. ### Claude + Plain Prompt ANLS: 0.0798. Result ID: 6991445a3a6c31484906e883. [Reporting paper]() Recorded evaluation configuration: {"Paradigm":"Zero-shot"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.