Patrick Wendell on Office QA benchmark
From How OpenAI and Databricks are working together · · Databricks
“And at any given time there's a leaderboard with you know who's doing the best. So so what makes Office QA a bit different as I said is it involves a lot of these fairly complex tasks around extracting data from documents and combining data across different sources and things that when you're working between applications answering a question running a process. Exactly. So it was designed exactly to capture those workloads because those are the ones that we at data bricks feel we're uniquely able to kind of explain as an enterprise company with you know a huge number of customers and um so whenever new models come out we sort of run them through the gauntlet and see how they do and uh and I'm happy to announce that as of yesterday when we did our sort of official run through of office QA with GPT 5.5 that 5.5 is now the leader state-of-the-art model for this particular workload.”
On , Patrick Wendell, Cofounder at Databricks, spoke about Office QA benchmark during How OpenAI and Databricks are working together on Databricks.
About Patrick Wendell
On April 24, 2026, Databricks co-founder Patrick Wendell appeared in a conversation with OpenAI CRO Denise Holland Dresser to discuss the launch of GPT-5.5 and the ongoing partnership between the two companies. Wendell described GPT-5.5 as a "really high quality model" and stated that it was his "most exciting model launch in a while." He announced that GPT-5.5 had become the "state-of-the-art model" for a particular workload after internal testing. Wendell shared anonymous user feedback describing the model as "a big step up from GPT 5.4" and noted its improved planning capabilities. Wendell discussed Databricks' focus on helping enterprise customers deploy AI agents, emphasizing the importance of both data and AI. He noted that while frontier models provide intelligence, accessing internal company data remains a challenge because "that data is messy" and "spread out." He also raised the issue of cost management, stating that "costs can get out of control once the use of AI tokens starts to really ramp." Wendell expressed optimism about expanding the use of GPT-5.5 beyond coding into other areas like marketing and data analytics, and said Databricks plans to make the model broadly available through its AI gateway product.