Allen Institute for Artificial Intelligence
PulseAugur coverage of Allen Institute for Artificial Intelligence — every cluster mentioning Allen Institute for Artificial Intelligence across labs, papers, and developer communities, ranked by signal.
- 2026-06-20 research_milestone Researchers developed a generative AI that emulates a climate model 25x faster, winning Best Paper at ICML 2024. source
14 day(s) with sentiment data
-
AI2 compares transformer and hybrid models on token processing
Researchers at AI2 compared their transformer model, Olmo 3, with a hybrid transformer-RNN model, Olmo Hybrid, to investigate differences in token processing and performance. The study aims to understand how these hybri…
-
Open-source OCR models and benchmarks consolidated on Papers with Code
A new resource has been created to track open-source optical character recognition (OCR) models, consolidating information on top-performing models, benchmarks, and links to their papers and code. This initiative highli…
-
Hugging Face releases new OCR and 3D motion models, plus AI website cloner
Hugging Face has released two new open-weight multimodal models: PP-OCRv6 for advanced OCR across 50 languages and MolmoMotion for language-guided 3D motion forecasting. These models are designed for accessibility, with…
-
AI accelerates climate modeling 25x faster, but with higher biases
Researchers from UC San Diego and the Allen Institute for AI have developed a generative AI capable of simulating climate models approximately 25 times faster than traditional methods. This AI can complete a 10-year cli…
-
AI Emulates Climate Model 25x Faster, Wins Best Paper at ICML 2024 · 2 sources tracked
Researchers from UC San Diego and the Allen Institute for Artificial Intelligence have developed a generative AI capable of emulating the NOAA FV3GFS climate model approximately 25 times faster. This AI can simulate 10 …
-
AI models frequently generate the same character, Elias Thorne, in stories
A recurring character named Elias Thorne, appearing as a clockmaker, lighthouse keeper, or librarian, is frequently generated by various large language models, including ChatGPT, Grok, Deepseek, and Gemini. Researchers …
-
Bluesky Jetstream enables custom AI models with open Olmo language models
Bluesky's Jetstream platform facilitated the development of custom AI models for regulated industries by thinkaisquared.bsky.social and Domyn. They utilized Olmo, a family of open language models, to build specialized m…
-
AI2 releases MolmoMotion for 3D motion forecasting
The Allen Institute for Artificial Intelligence (AI2) has released MolmoMotion, a new 3D motion forecasting model. This model can predict the future movement of 3D points on an object based on a few video frames and an …
-
AI Analyst Nathan Lambert Reframes Blog for Open Science Mission
Nathan Lambert, formerly of the Allen Institute for Artificial Intelligence, is redefining the focus of his blog, Interconnects. He aims to provide clarity on frontier AI models, foster an open model ecosystem, and buil…
-
LLM post-training recipes evolve with new distillation techniques
A review of post-training recipes for large language models highlights significant evolution in the past year. Historically, models followed a pipeline of Supervised Fine-Tuning (SFT), reward modeling, and Reinforcement…
-
AI Researchers Adopt Automated Tools for Faster Model Evaluation
AI researchers are adopting automated workbenches to accelerate model evaluation, moving away from slower manual methods. A notable new tool is olmo eval, developed by AI2, which aims to streamline the model development…
-
Hugging Face Spotlights AI Tools: Holotron-12B, olmo-eval, and PyTorch Profiler
Hugging Face is highlighting several AI-related projects and tools. This includes Holotron-12B, an agent designed for high-throughput computing, and olmo-eval, an evaluation workbench for model development loops from th…
-
AI Labs Announce New Models, Courses, and Evaluation Tools
Several AI-focused entities have released new information or initiatives. Google DeepMind has announced DiffusionGemma, a model capable of generating text four times faster. OpenAI is launching new courses through its O…
-
AI2 releases olmo-eval for iterative LLM development
The Allen Institute for Artificial Intelligence (AI2) has released olmo-eval, a new workbench designed to streamline the iterative evaluation process required when building Large Language Models (LLMs). This tool aims t…
-
New Arabic LLM Leaderboard and Earth Observation Models Released
The QIMMA leaderboard has been released, focusing on the quality of Arabic Large Language Models (LLMs). Separately, Allen Institute for AI has launched OlmoEarth v1.1, a collection of more efficient models designed for…
-
AI2 presents research at CVPR 2026 conference
AI2 is participating in the CVPR 2026 conference, presenting research through papers and talks. Attendees are invited to visit their booth to learn more about their latest work.
-
Nathan Lambert departs Allen Institute for AI, cites open ecosystem focus
Nathan Lambert is departing the Allen Institute for AI (Ai2) after a two-year tenure where he contributed to projects like the Olmo models. He reflects on Ai2's influential role in the AI research community, balancing a…
-
Allen Institute for AI project contributor departs
Nathan Lambert has concluded his tenure at the Allen Institute for AI (Ai2). During his over two years with the organization, he contributed to significant projects such as Olmo and Tulu. Lambert expressed gratitude for…
-
SERA method trains specialized coding agents cheaply
Researchers have developed SERA, a method for efficiently training coding agents specialized for private codebases. This approach uses Soft Verified Generation to create synthetic training data without requiring unit te…
-
Bluesky AI2 releases open robotics model code and data
Bluesky's AI2 team has released the full code and training data for their robotics foundation model, MolmoAct 2. The model's artifacts have already seen over 400,000 downloads in less than a month. This release aims to …