APEX-Agents: AI agent benchmark
480 tasks from bankers, consultants, and lawyers to test AI agents in cross-application scenarios.
- English
Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,788 in the catalog so far, with new ones added every hour.
Answered by Gemini 4 · the chat is kept only in this tab
480 tasks from bankers, consultants, and lawyers to test AI agents in cross-application scenarios.
Unedited daily and working first-person video recordings in English for training video models.
Multimodal corpus of parametric 3D models from Objaverse featuring sand3d code, OpenUSD scenes, and images.
8,189 flower images with annotations for classification and segmentation in the FiftyOne library.
10 to 100 thousand images to test model robustness against visual noise.
35,000 food photographs for image classification in the FiftyOne library format.
Over 5 thousand Chinese phraseological units with text explanations and generated images for multimodal tasks.
A dataset based on reviews and author responses to evaluate text and visual analysis of papers.
From 1 to 10 thousand examples for evaluating robotics VLA models in simulation.
90,289 raw GoPro photos for 3D photogrammetry and coral reef reconstruction.
A mix of expert RL demonstrations, image-caption pairs, and text for training a multimodal agent.
A collection of 50 training vision-language datasets used for fine-tuning Idefics2.
A three-configuration set featuring CadQuery 3D code, part renders, editing tasks, and CAD QA.
Generated images, genomes, and VLM agent reasoning logs from the Picbreeder study.
Over 8,100 English and Chinese prompts together with generated videos and preview frames.
Data from a 128-channel LiDAR, stereo cameras, and a navigation system for depth estimation tasks.
Raw data, images, and English text for anomaly detection and segmentation in 3D printing.
A synthetic counterfactual corpus of images and scene graphs to study numerals in VLMs.
A balanced set of JPEG hand gesture images for training classification and detection models.
Multimodal set of 4,307 characters with AI images, definitions, and pinyin audio for generative and language models.
First dataset with images, text, and time series for semantic communication and 6G tasks.
A huge corpus of multimodal instructions (image + text) for full-parameter SFT of vision-language models.
19,200 JPEG images with 24 hand gesture classes, split into train, validation, and test sets.
A dataset of baggage X-ray images with bounding box and segmentation mask annotations for 12 categories of hidden contraband.