[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

[@allen_ai](/creator/twitter/allen_ai)
"Introducing olmOCR our open-source tool to extract clean plain text from PDFs Built for scale olmOCR handles many document types with high throughput. Run it on your own GPU for freeat over 3000 token/s equivalent to $XXX per million pages or 1/32 the cost of GPT-4o"  
[X Link](https://x.com/allen_ai/status/1894415487569969211) [@allen_ai](/creator/x/allen_ai) 2025-02-25T15:53Z 74.1K followers, 284.9K engagements


"📣 Heads up Bay Area folks: Research Scientists Nathan Lambert (@natolambert) & Sewon Min (@sewon__min) will be giving separate talks during #OpenSourceAIWeek and #PyTorchCon in SF"  
[X Link](https://x.com/allen_ai/status/1980695820593684863) [@allen_ai](/creator/x/allen_ai) 2025-10-21T18:00Z 74.1K followers, 2555 engagements


"Introducing Asta DataVoyagerour new AI capability in Asta that turns structured data into transparent reproducible insights. Built for scientists grounded in open inspectable workflows. 🧵"  
[X Link](https://x.com/allen_ai/status/1973372657618870461) [@allen_ai](/creator/x/allen_ai) 2025-10-01T13:01Z 74.2K followers, 368.1K engagements


"@natolambert @sewon__min 📅 Thurs Oct XX XX PM PDT 📍 UC Law SF Join An Evening of Open: Science Software and AI co-hosted by @UCLaw_SF @GitHubPolicy @FranceinSF. @sewon__min joins Powering the Future of Research. RSVP:"  
[X Link](https://x.com/allen_ai/status/1980695823974285732) [@allen_ai](/creator/x/allen_ai) 2025-10-21T18:00Z 74.2K followers, 1609 engagements


"On olmOCR-Bench olmOCR X scores XXXX points up from XXXX in our previous releaseincreasing performance across every document category. 📈"  
[X Link](https://x.com/allen_ai/status/1981029163394797618) [@allen_ai](/creator/x/allen_ai) 2025-10-22T16:05Z 74.2K followers, 157.2K engagements


"We now ship an FP8 quantized model (olmOCR-2-7B-1025-FP8) that gets 3400 output tokens/sec on a single H100. 🚀 That means you can process 10000 pages for $X. Want to fine-tune olmOCR X on your own data We provide everything you need for that too. ⚡"  
[X Link](https://x.com/allen_ai/status/1981029164615356755) [@allen_ai](/creator/x/allen_ai) 2025-10-22T16:05Z 74.2K followers, 6534 engagements


"Were updating our demo and olmOCR-2-7B-1025 is now available on @DeepInfra & @parasail_io APIs. Check out our repos for instructions on running locally. 📚 Learn more: 💻 Model: 💬 Discuss:"  
[X Link](https://x.com/allen_ai/status/1981029165517177257) [@allen_ai](/creator/x/allen_ai) 2025-10-22T16:05Z 74.2K followers, 6957 engagements