Today we’re announcing r-1, our new document parsing model. It’s more accurate than our most powerful agentic OCR models, faster, and up to 6x cheaper. At @reductoai, we spent two years building specialized models for complex visual layouts, tables spanning multiple pages, and key formatting like strikethroughs. We then used everything we learned to build r-1, the first in a new generation of models designed to handle the hardest documents without multiplying cost. This early preview delivers a 20% lower error rate than our most accurate legacy agentic models and will keep improving with new checkpoints over the next few weeks. Accuracy is only the beginning. r-1 preview is available at 1¢ per page all-in, with additional volume discounts as you scale. In the near term we’re also going to release r-1 mini, and an auto mode that intelligently selects the right approach for each page. If you’re currently using another parser, we’re offering up to $5,000 in credits to evaluate and migrate. You can claim the migration offer and learn more about r-1 using the links in the comments. Happy parsing!

By “a lot” I mean enough to actually push for the outcome you want. Caring enough to want something but not enough to do anything meaningful for it is frustrating.

You should either care a lot or choose to not care. Caring a medium amount is the worst.

Life changed when I learned that uptown trains go up on the map and downtown trains go down on the map (derived from first principles)

A list of talent with great taste whose range can’t be captured in a single label cosign.co/lists/eriktorenber…

we're taking over NYC this week! come say hi if you're at @aiDotEngineer 🗽 find us at booth 34 in the Empire Ballroom all 3 days 🔮

You come across a puddle that’s 1.75x your stride in length. Do you: - Do a dainty lil hop - Stay nonchalant and step into it - Find a new route

I don’t think I’ve ever responded to a cold email where the ask was a meeting to “trade notes”

Can a tire expert confirm this can be patched

I’m just a ramp card to Adel

what does "good" document processing even mean? benchmarking in this space can mean a lot of things: > accuracy on complex docs > latency > throughput for scale we're not here to tell you what set of documents is a "definitive benchmark", because your own docs are the ones that matter. the more important part is actually deciding how you measure performance. @archanaahlawat and @jannalulu_ have been working a lot on building and tuning our in-house judge framework, and turns out there's a couple of reoccurring patterns we see that might help you with your eval process:

some shots from our @reductoai offsite.....can't wait to look through all the video footage 😍 i'm kind of @aditabrm's personal paparazzi

Accurate, fast, and priced to scale

I love seeing charts like this because they're a great reminder to zoom out Even if this is just normal seasonality for Devin, when you're actually in a period like this it's really easy to think short term

Whole team is staying in Pablo Escobar's mansion with a private beachfront next week and Sleepy Donald is upset he has to set his alarm a little early 😴