Ai2
@ai2.bsky.social
Breakthrough AI to solve the world's biggest problems.
› Join us: http://allenai.org/careers
› Get our newsletter: https://share.hsforms.com/1uJkWs5aDRHWhiky3aHooIg3ioxm
› Join us: http://allenai.org/careers
› Get our newsletter: https://share.hsforms.com/1uJkWs5aDRHWhiky3aHooIg3ioxm
Pinned
Ai2
@ai2.bsky.social
· 9d
Introducing OlmoEarth 🌍, state-of-the-art AI foundation models paired with ready-to-use open infrastructure to turn Earth data into clear, up-to-date insights within hours—not years.
Introducing OlmoEarth 🌍, state-of-the-art AI foundation models paired with ready-to-use open infrastructure to turn Earth data into clear, up-to-date insights within hours—not years.
November 4, 2025 at 2:52 PM
Introducing OlmoEarth 🌍, state-of-the-art AI foundation models paired with ready-to-use open infrastructure to turn Earth data into clear, up-to-date insights within hours—not years.
Our Olmo Discord AMA just wrapped! Researchers answered community questions about their work with Olmo, our family of fully open LLMs. Here’s some highlights. 🧵
October 28, 2025 at 6:56 PM
Our Olmo Discord AMA just wrapped! Researchers answered community questions about their work with Olmo, our family of fully open LLMs. Here’s some highlights. 🧵
⚠️ Reminder: Join us today in our Discord @ 8:00 a.m. PT!
Join us for a live Discord AMA Tues, Oct 28 @ 8:00 a.m. PT with researchers who’ve run real studies using our Olmo language model family—from machine unlearning to knowledge cutoffs to how models acquire new skills. 👇
October 28, 2025 at 1:17 PM
⚠️ Reminder: Join us today in our Discord @ 8:00 a.m. PT!
Join us for a live Discord AMA Tues, Oct 28 @ 8:00 a.m. PT with researchers who’ve run real studies using our Olmo language model family—from machine unlearning to knowledge cutoffs to how models acquire new skills. 👇
October 27, 2025 at 4:01 PM
Join us for a live Discord AMA Tues, Oct 28 @ 8:00 a.m. PT with researchers who’ve run real studies using our Olmo language model family—from machine unlearning to knowledge cutoffs to how models acquire new skills. 👇
Our fully open Olmo models enable rigorous, reproducible science—from unlearning to clinical NLP, math learning, & fresher knowledge. Here’s how the research community has leveraged Olmo to make the entire AI ecosystem better + more transparent for all. 🧵
October 24, 2025 at 6:36 PM
Our fully open Olmo models enable rigorous, reproducible science—from unlearning to clinical NLP, math learning, & fresher knowledge. Here’s how the research community has leveraged Olmo to make the entire AI ecosystem better + more transparent for all. 🧵
Curious to see how olmOCR 2, our new tool to turn digitized documents into trustworthy, structured text, performs? Check out these examples. 👇
October 22, 2025 at 9:20 PM
Curious to see how olmOCR 2, our new tool to turn digitized documents into trustworthy, structured text, performs? Check out these examples. 👇
We’re updating olmOCR, our model for turning PDFs & scans into clean text with support for tables, equations, handwriting, & more. olmOCR 2 uses synthetic data + unit tests as verifiable rewards to reach state-of-the-art performance on challenging documents. 🧵
October 22, 2025 at 4:09 PM
We’re updating olmOCR, our model for turning PDFs & scans into clean text with support for tables, equations, handwriting, & more. olmOCR 2 uses synthetic data + unit tests as verifiable rewards to reach state-of-the-art performance on challenging documents. 🧵
📣 Heads up, Bay Area folks: Research Scientists Nathan Lambert (@natolambert.bsky.social) & Sewon Min (@sewonm.bsky.social) will be giving separate talks during #OpenSourceAIWeek and #PyTorchCon in SF.
October 21, 2025 at 6:01 PM
📣 Heads up, Bay Area folks: Research Scientists Nathan Lambert (@natolambert.bsky.social) & Sewon Min (@sewonm.bsky.social) will be giving separate talks during #OpenSourceAIWeek and #PyTorchCon in SF.
🌍 Announcing SamudrACE, our AI climate emulator built so scientists & planners can run “what-if” climate experiments quickly. Traditional models are slow and costly; SamudrACE makes high-quality simulations fast & more accessible. 🧵
October 16, 2025 at 3:05 PM
🌍 Announcing SamudrACE, our AI climate emulator built so scientists & planners can run “what-if” climate experiments quickly. Traditional models are slow and costly; SamudrACE makes high-quality simulations fast & more accessible. 🧵
We were honored to join the #IUCNcongress 2025 in Abu Dhabi last week. Our message: AI should serve science & society—built openly, deployed with partners, and measured by real-world impact.
October 13, 2025 at 8:25 PM
We were honored to join the #IUCNcongress 2025 in Abu Dhabi last week. Our message: AI should serve science & society—built openly, deployed with partners, and measured by real-world impact.
📊 Today we're releasing data showing which scientific papers our AI research tool Asta cites most frequently. Think of it as creating citation counts for the AI era—tracking which research is actually powering AI answers across thousands of queries. 🧵
October 8, 2025 at 6:26 PM
📊 Today we're releasing data showing which scientific papers our AI research tool Asta cites most frequently. Think of it as creating citation counts for the AI era—tracking which research is actually powering AI answers across thousands of queries. 🧵
"We are committed to our fully open ethos. That's why we release everything—weights, code, training data, checkpoints, all of it." — @nlpnoah.bsky.social at the Madrona IA Summit last week.
October 6, 2025 at 5:46 PM
"We are committed to our fully open ethos. That's why we release everything—weights, code, training data, checkpoints, all of it." — @nlpnoah.bsky.social at the Madrona IA Summit last week.
As part of #SeattleAIWeek, we're hosting "AI Innovation in the Open" on Oct. 30 from 2-4:30pm—an afternoon of live demos and hands-on tutorials at Ai2 HQ. 👇
October 2, 2025 at 7:06 PM
As part of #SeattleAIWeek, we're hosting "AI Innovation in the Open" on Oct. 30 from 2-4:30pm—an afternoon of live demos and hands-on tutorials at Ai2 HQ. 👇
Introducing Asta DataVoyager—our new AI capability in Asta that turns structured data into transparent, reproducible insights. Built for scientists, grounded in open, inspectable workflows. 🧵
October 1, 2025 at 1:02 PM
Introducing Asta DataVoyager—our new AI capability in Asta that turns structured data into transparent, reproducible insights. Built for scientists, grounded in open, inspectable workflows. 🧵
A few new challengers enter SciArena—including DeepSeek-V3.2-Exp and Claude Sonnet 4.5 🔬
September 29, 2025 at 7:20 PM
A few new challengers enter SciArena—including DeepSeek-V3.2-Exp and Claude Sonnet 4.5 🔬
Reposted by Ai2
📢 New #COLM2025 paper 📢
Standard benchmarks give every LLM the same questions. This is like testing 5th graders and college seniors with *one* exam! 🥴
Meet Fluid Benchmarking, a capability-adaptive eval method delivering lower variance, higher validity, and reduced cost.
🧵
Standard benchmarks give every LLM the same questions. This is like testing 5th graders and college seniors with *one* exam! 🥴
Meet Fluid Benchmarking, a capability-adaptive eval method delivering lower variance, higher validity, and reduced cost.
🧵
September 16, 2025 at 5:16 PM
📢 New #COLM2025 paper 📢
Standard benchmarks give every LLM the same questions. This is like testing 5th graders and college seniors with *one* exam! 🥴
Meet Fluid Benchmarking, a capability-adaptive eval method delivering lower variance, higher validity, and reduced cost.
🧵
Standard benchmarks give every LLM the same questions. This is like testing 5th graders and college seniors with *one* exam! 🥴
Meet Fluid Benchmarking, a capability-adaptive eval method delivering lower variance, higher validity, and reduced cost.
🧵
🚀 Introducing Fluid Benchmarking—an adaptive way to evaluate LLMs. Inspired by psychometrics, it tailors which questions to ask based on each model’s capability, making evals more efficient & reliable. 🧵
September 16, 2025 at 4:08 PM
🚀 Introducing Fluid Benchmarking—an adaptive way to evaluate LLMs. Inspired by psychometrics, it tailors which questions to ask based on each model’s capability, making evals more efficient & reliable. 🧵
📓 New from Ai2: we’ve released source code that shows how we built AskOlmo, our Discord chatbot powered by our Olmo model family.
It’s a peek behind the curtain—so you can see how it all came together. 👇
It’s a peek behind the curtain—so you can see how it all came together. 👇
September 10, 2025 at 7:45 PM
📓 New from Ai2: we’ve released source code that shows how we built AskOlmo, our Discord chatbot powered by our Olmo model family.
It’s a peek behind the curtain—so you can see how it all came together. 👇
It’s a peek behind the curtain—so you can see how it all came together. 👇
🚀 New in the Ai2 Playground: side-by-side comparison is live.
Compare two Ai2 models with the same prompt and see the results next to each other. ⚖️🆚
Compare two Ai2 models with the same prompt and see the results next to each other. ⚖️🆚
September 4, 2025 at 6:03 PM
🚀 New in the Ai2 Playground: side-by-side comparison is live.
Compare two Ai2 models with the same prompt and see the results next to each other. ⚖️🆚
Compare two Ai2 models with the same prompt and see the results next to each other. ⚖️🆚
🌍☀️❄️ Can AI forecast year-to-year differences in the seasons?
New research with @metoffice.gov.uk shows our ACE2 ML model demonstrates seasonal forecasting skill—matching traditional physics-based methods while using dramatically less compute. 🧵
New research with @metoffice.gov.uk shows our ACE2 ML model demonstrates seasonal forecasting skill—matching traditional physics-based methods while using dramatically less compute. 🧵
September 4, 2025 at 2:04 PM
🌍☀️❄️ Can AI forecast year-to-year differences in the seasons?
New research with @metoffice.gov.uk shows our ACE2 ML model demonstrates seasonal forecasting skill—matching traditional physics-based methods while using dramatically less compute. 🧵
New research with @metoffice.gov.uk shows our ACE2 ML model demonstrates seasonal forecasting skill—matching traditional physics-based methods while using dramatically less compute. 🧵
Reposted by Ai2
✨Meet OLMoASR✨ By pairing our curated 1M-hour dataset with a powerful architecture, we've built open ASR models that achieve competitive performance with models like Whisper. We're open-sourcing data, code and models to help the community build more robust and transparent ASR.
🎙️ Say hello to OLMoASR—our fully open, from-scratch speech-to-text (STT) model. Trained on a curated audio-text set, it boosts zero-shot ASR and now powers STT in the Ai2 Playground. 👇
August 29, 2025 at 4:21 PM
✨Meet OLMoASR✨ By pairing our curated 1M-hour dataset with a powerful architecture, we've built open ASR models that achieve competitive performance with models like Whisper. We're open-sourcing data, code and models to help the community build more robust and transparent ASR.
🎙️ Say hello to OLMoASR—our fully open, from-scratch speech-to-text (STT) model. Trained on a curated audio-text set, it boosts zero-shot ASR and now powers STT in the Ai2 Playground. 👇
August 28, 2025 at 4:13 PM
🎙️ Say hello to OLMoASR—our fully open, from-scratch speech-to-text (STT) model. Trained on a curated audio-text set, it boosts zero-shot ASR and now powers STT in the Ai2 Playground. 👇
We're honored to be named a @fastcompany.com Corporate Wellness Innovator 2025 (Tech)! 🏆 Our Seattle HQ was built for well-being with fresh air, outdoor spaces, and movement-friendly designs to help our team thrive.
August 27, 2025 at 7:58 PM
We're honored to be named a @fastcompany.com Corporate Wellness Innovator 2025 (Tech)! 🏆 Our Seattle HQ was built for well-being with fresh air, outdoor spaces, and movement-friendly designs to help our team thrive.
Today we’re releasing agent-baselines, a suite of 22 classes of AI agents for science—including 9 open-source research-tuned agents like our state-of-the-art, benchmark-leading Asta v0. 🚀🔬
Part of our Asta ecosystem to advance scientific AI. 👇
Part of our Asta ecosystem to advance scientific AI. 👇
August 26, 2025 at 7:45 PM
Today we’re releasing agent-baselines, a suite of 22 classes of AI agents for science—including 9 open-source research-tuned agents like our state-of-the-art, benchmark-leading Asta v0. 🚀🔬
Part of our Asta ecosystem to advance scientific AI. 👇
Part of our Asta ecosystem to advance scientific AI. 👇
As part of Asta, our initiative to accelerate science with trustworthy AI agents, we built AstaBench—the first comprehensive benchmark to compare them. ⚖️
August 26, 2025 at 3:02 PM
As part of Asta, our initiative to accelerate science with trustworthy AI agents, we built AstaBench—the first comprehensive benchmark to compare them. ⚖️