CLARU
RLE-Bench

kinex (v0.6.0) vs kinex (v0.6.0-improve) vs codex · LIBERO Long · GPT-6 Astra / medium

Evaluation dataset comparing three agent variants (kinex v0.6.0, kinex v0.6.0-improve, and codex) on LIBERO Long tasks using GPT-6 Astra medium model with 45 planned episodes.

Downloads673
Episodes45

Why This Matters for Physical AI

This evaluation dataset enables comparative analysis of different embodied AI agent implementations on standardized manipulation benchmarks, supporting research into agent design improvements and model capabilities for robotic task execution.

Technical Profile

Episodes
45
Annotation Types
action_labelsreward_labels
Part of the LIBERO family

Access

View on HuggingFace

Need custom physical AI data?

Claru builds purpose-built datasets for any environment applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets

TCoT LIBERO RLDS

RLDS/TFDS training data for TCoT experiments containing four filtered successful-demonstration suites and terminal windows of failed/retry rollouts from libero_10_end60_retry.

746 downloadsAug 2026other

LIBERO LeRobot v3.0

Four standard LIBERO demonstration suites packaged in LeRobot v3.0 format containing synchronized RGB observations, robot state, actions, and natural-language task metadata for robot-policy and world-action-model training.

rgbproprioceptionlanguage
191 downloadsSep 2026cc-by-4.0

LIBERO baseline rollout videos — pi0 / pi0-FAST

Rendered rollout videos from lerobot-eval runs of pi0/pi0-FAST checkpoints on the LIBERO benchmark, featuring a Franka Panda arm with task instructions and success/failure labels.

rgb
132 downloadsSep 2026apache-2.0

ActionVAE processed training sidecars

Compact processed artifacts including paired LIBERO z-effect/z-action teacher caches, RoboTwin paired teacher caches, ActionAE sidecar protocols, and compressed LARYBench pilot archives for offline validation used in ActionVAE and phased LaWAM training pipelines.

105 downloadsSep 2026other

LIBERO

A vision-language-action dataset for robot manipulation with discrete actions, organized into spatial, object, goal, and long-horizon task categories.

rgblanguage
93 downloadsSep 2026MITgated

LIBERO-Plus_custom

Custom multi-view dataset composed of simulation scenes from LIBERO-plus assets, rendered from multiple viewpoints for finetuning the LVSM novel view synthesis model used in AnyCamVLA.

rgb
64 downloadsAug 2026MIT License (upstream LIBERO)