Allen AI
Allen Institute for AI, fully-open models (OLMo) with training data, code, and weights released together.
Release history
Rslearn 0.1.18
V1.3 renaming
Design 0.3.0
Internal design-system tooling update from Allen AI; Limited relevance outside their stack
Asta Plugins 0.109.0
Adds Overleaf pull/publish support and fork-PR handling
Rslearn 0.1.17
Rslearn adds Sentinel-3 OLCI/SLSTR sources, ECOSTRESS data, per-pixel timestep prediction, and fixes image formatting bugs
Vla Evaluation Harness 0.8.0
Shared episode scheduling; Clearer shard failures; More reliable RoboTwin and container execution
Rslearn 0.1.16
Loads released bundle with full olmoearth pretrain loader and supports overriding ground control point handling
Rslearn 0.1.15
Makes WorldCereal downloads resumable with checksum verification, adds sdpa option, and selects distilled register student for embedding extraction
Vla Evaluation Harness 0.7.0
More reliable recording and concurrent container runs, with CPU rendering for additional benchmarks
Olmoearth Pretrain Minimal 0.0.9
Fixes inconsistent CompositeEncodings pos embed handling
Molmospaces 0.2.9
Stable release tagged before major update
Kernel Fun 0.2.0
First PyPI upload; V0.2.0.dev0 workflow could not resolve setup-uv so build never ran
Asta Autodiscovery 1.0.1
Published to PyPI: asta-autodiscovery 1.0.1
Asta Autodiscovery 1.0.0
Published to PyPI: asta-autodiscovery 1.0.0
Rslearn 0.1.14
Enhances EarthDaily integration with asset key resolution, per-pixel regression metrics bugfix, logger config saving fix
Design.Md 0.2.0
Design.Md now records OlmoEarth subtitle roles and prevents duplicate version tags during releases
Design.Md 0.1.0
Design.md publishes Skylight and EarthRanger design systems with agent skills framework to npm
Olmoearth Pretrain Minimal Publish 0.0.7
Removed max_sequence_length constraint to enable flexible sequence handling in pretraining pipeline
Asta Bench Leaderboard Design review attachments
Rendered review evidence hosted outside the source tree and referenced from pull-request metadata
Vla Evaluation Harness 0.4.0
Recording on; LeRobot bridge ships with four reproduced LIBERO scores and reproducible Docker builds