The Best Model Wont Stay The Best @Databricks
The Best Model Wont Stay The Best  @Databricks
Uploaded November 2026 | Updated November 2026, 1 day ago
Are you paying too much for LLMs? In this talk, Yuchen Jin (AI Inference Lead at Databricks) breaks down why the intelligence gap between closed-source frontier models and open-source AI is rapidly closing—and why relying on public benchmarks like SWE-bench or static leaderboards can mislead enterprise AI strategies.

Discover how top engineering teams evaluate large language models (LLMs) using Pareto frontier curves (accuracy vs. cost per task), optimize AI inference performance, and build custom evaluation frameworks (Evals) to drive lower costs and higher speed.

Explore more at: databricks.com/product/artificial-intelligence/unity-gateway
The Best Model Wont Stay The BestLTAP Explained:  How Databricks Unifies OLTP and OLAPDatabricks CustomerLakeHow Suntory Turns Data into Faster Decisions with DatabricksDemo: Building a Governed AI Agent with Unity AI GatewayDatabricks AI Functions: 3-5x cheaper than frontier modelsWhat’s coming next to Free EditionAli Ghodsi, Co-founder and CEO, Databricks kicks off Data + AI Summit 2026When to choose CPU vs GPU: Databricks AI Runtime ExplainedBrand new #data and #ai features in #databricks with ZerobusBuilding Trustworthy, High-Quality AI Agents with MLflowInside Lakebase: fully-managed serverless Postgres – Nikita Shamgunov, VP, Engineering, Databricks
Databricks |

The Best Model Won't Stay The Best

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER