A test run is a single execution of a benchmark test using a defined model configuration.
Each run represents how a particular large language model (LLM) — such as GPT-4, Claude-3, or Gemini — performed on a given task at a specific time, with specific settings.
A test run includes:
Together, test runs make it possible to compare models, providers, and configurations across benchmarks in a transparent and reproducible way.
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | openrouter |
| Model | z-ai/glm-5v-turbo |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 0.00 % |
| Test time | 24.92 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[{"box":[281.0,58.0,676.0,462.0]},{"box":[68.0,494.0,470.0,904.0]},{"box":[489.0,494.0,885.0,906.0]}]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 186.1K IT + 28.8K OT = 214.9K TT | Cost: 0.223$ + 0.115$ = 0.339$ |
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | genai |
| Model | gemini-3.7-flash |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 70.50 % |
| Test time | 5.56 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[{"box":[185.0,339.0,2155.0,1723.0]},{"box":[186.0,1780.0,2155.0,3144.0]}]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 52.3K IT + 5.1K OT = 57.4K TT | Cost: 0.039$ + 0.019$ = 0.058$ |
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | x-ai |
| Model | grok-4.6 |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 2.30 % |
| Test time | 1.80 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[{"box":[12.6,8.6,95.6,97.6]}]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 130.6K IT + 2.7K OT = 133.3K TT | Cost: 0.261$ + 0.016$ = 0.277$ |
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | openrouter |
| Model | meta/muse-spark-1.2 |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 87.80 % |
| Test time | 10.84 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[{"box":[1405.0,2465.0,2385.0,3185.0]}]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 177.4K IT + 114.9K OT = 292.3K TT | Cost: 0.222$ + 0.488$ = 0.710$ |
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | huggingface |
| Model | meta-models/Muse-Glimmer-30B |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 39.60 % |
| Test time | 260.63 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[{"box":[105.0,60.0,2380.0,260.0]},{"box":[105.0,275.0,1220.0,480.0]},{"box":[1320.0,275.0,2380.0,480.0]},{"box":[105.0,495.0,1220.0,700.0]},{"box":[1320.0,495.0,2380.0,700.0]},{"box":[105.0,715.0,1220.0,920.0]},{"box":[1320.0,715.0,2380.0,920.0]}]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 180.1K IT + 54.6K OT = 234.7K TT | Cost: 0.063$ + 0.082$ = 0.145$ |
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | huggingface |
| Model | thinkingmachines/Inkling-Small |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 53.50 % |
| Test time | 4.95 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[{"box":[0.36,0.17,0.82,0.37]}]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 91.2K IT + 49.3K OT = 140.5K TT | Cost: 0.041$ + 0.059$ = 0.100$ |
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | huggingface |
| Model | thinkingmachines/Inkling |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 13.50 % |
| Test time | 46.35 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[{"box":[90.0,190.0,2390.0,540.0]},{"box":[1440.0,570.0,2380.0,1010.0]},{"box":[1440.0,1070.0,2380.0,2130.0]}]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 92.4K IT + 34.7K OT = 127.1K TT | Cost: 0.092$ + 0.141$ = 0.233$ |
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | huggingface |
| Model | Qwen/Qwen3-VL-235B-A22B-Instruct |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 0.00 % |
| Test time | 5.88 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[{"box":[279.0,60.0,678.0,460.0]},{"box":[67.0,490.0,467.0,907.0]},{"box":[484.0,490.0,884.0,907.0]}]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 135.1K IT + 4.0K OT = 139.2K TT | Cost: 0.027$ + 0.004$ = 0.031$ |
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | huggingface |
| Model | MiniMaxAI/MiniMax-M3 |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 12.30 % |
| Test time | 2.08 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[{"box":[137.0,86.0,1100.0,1370.0]}]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 35.4K IT + 2.5K OT = 37.9K TT | Cost: 0.010$ + 0.003$ = 0.013$ |
newspaper-page document-understanding printed latin 20 en prose, columns
| Provider | huggingface |
| Model | swiss-ai/apertus-v1.5-70b |
| Temperature | 0.0 |
| Dataclass | MagazinePage |
| Normalized Score | 0.00 % |
| Test time | 2.79 seconds |
Extract all advertisements and return their bounding boxes.
The original size of the page is {width} x {height} pixels.
{"advertisements":[]}| Fuzzy Score | F1 micro / macro | Micro precision/recall | Tue/False Positives | |||||
| n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a | n/a |
| Micro Precision | Micro Recall | Instances | TP | FP | FN | |||
| Pricing Date: n/a, n/a. | Tokens: 362.1K IT + 3.5K OT = 365.6K TT | Cost: 0.297$ + 0.010$ = 0.307$ |