1,818 submissions later, Dr.DocBench gave us plenty to think about:
📖 Fiction pages? The top 20 averaged 99.9.
🧪 Page structure? On a page combining a table and a long list, all top-20 systems scored below 50.
🌿 A plant index? Two teams reached 86.66; the other 18 scored
We're throwing the swankiest event of #CORL2026.
This night is for the jazz lovers, the smooth operators, and the bourbon sippers.
Date: Tues, Nov 10
Time: 19:00–23:00
Address: Secret location, 3-min walk from the JW Marriott (Austin, TX)
Space is limited. RSVP below.
Wanna watch startup founders get roasted?
SF-based comedian Cathy Chao is doing a night of live stand-up about AI, startups, and Silicon Valley at our office in Mountain View.
Date: Thursday, Oct 29
Time: 18:30–21:00 (show starts at 19:00)
RSVP below ⬇️
Dr.DocBench closes in 3 days, with a $5,000+ prize pool, together with DocSem.
How far can your parser climb on the leaderboard?
Submit by October 10 at 12:59 PM UTC.
Try the challenge: drdocbench-challenge.abaka-pages.com#EMNLP2026
12 parsers and VLMs tested, with 61.94 as the best overall score.
Try your model on expert-level document parsing in the Dr.DocBench Challenge at DocInsights.
Submit by October 10, 12:59 UTC. Details below.
#EMNLP2026
