Library
85,340 resources from 11 public sources. 54 checked by hand; the rest are unreviewed.
Results
-
PhysicalAI-Autonomous-Vehicles
Datasets & benchmarksHugging Face dataset by nvidia
nvidia/PhysicalAI-Autonomous-Vehicles -
PhysicalAI-Autonomous-Vehicles-NuRec
Datasets & benchmarksHugging Face dataset by nvidia
nvidia/PhysicalAI-Autonomous-Vehicles-NuRec -
PhysicalAI-Robotics-GR00T-X-Embodiment-Sim
Datasets & benchmarksHugging Face dataset by nvidia
nvidia/PhysicalAI-Robotics-GR00T-X-Embodiment-Sim -
physics-problems
Datasets & benchmarksHugging Face dataset by HumynLabs
HumynLabs/physics-problems -
PhysGym
Datasets & benchmarksA benchmark suite for evaluating LLM-based interactive scientific reasoning.
git clone https://github.com/principia-ai/PhysGym.git -
Open-RL
Datasets & benchmarksHugging Face dataset by TuringEnterprises
TuringEnterprises/Open-RL -
Turing-Open-Reasoning
Datasets & benchmarksHugging Face dataset by TuringEnterprises
TuringEnterprises/Turing-Open-Reasoning -
DishwasherBench
Datasets & benchmarksA dishwasher-loading game on a physical simulation of one real machine, built to benchmark language models and playable in a browser. The game inside is called Spotless.
git clone https://github.com/composo-ai/DishwasherBench.git -
Capybara
Datasets & benchmarksHugging Face dataset by LDJnr
LDJnr/Capybara -
Open-MM-RL
Datasets & benchmarksHugging Face dataset by TuringEnterprises
TuringEnterprises/Open-MM-RL -
peS2o
Datasets & benchmarksHugging Face dataset by allenai
allenai/peS2o -
seed
Datasets & benchmarksHugging Face dataset by bones-studio
bones-studio/seed -
distilabel-capybara-dpo-7k-binarized
Datasets & benchmarksHugging Face dataset by argilla
argilla/distilabel-capybara-dpo-7k-binarized -
Awesome-Scientific-LLM-Benchmarks
Datasets & benchmarksA curated, accuracy-first list of benchmarks for evaluating LLMs on scientific reasoning and discovery — math, physics, chemistry, materials, biology, and agentic science.
git clone https://github.com/subinium/Awesome-Scientific-LLM-Benchmarks.git
Indexed entries aren't reviewed or tested by InnerVault. Details come straight from each source. Anything that can reach your accounts, files or browser can do damage, so read the source before you install it.