Research to Advance AI
Scale Labs advances AI through research. Our research focuses on agents, post-training, reasoning, safety, evaluation, and alignment, and the science of data.
[LEADERBOARDS]
Benchmarks for frontier, agentic, and safety capabilities
[PAPERS]
Research papers and publications covering agents, post-training, reasoning, safety, evaluation, and alignment, and the science of data.
SteerDuplex: Steerable Duplex Speech Dialogue Models
[BLOG]
Insights, analysis, and updates from Scale Labs
SWE-Bench Pro V2: A Cleaner, Harder-to-Game Leaderboard
When SWE Bench Pro originally launched a little over a year ago, it was a significant entry into the open source benchmark space and changed how the field measured coding agents. However, we know that benchmarks built on open source GitHub issues have real limitations. We see that as a reason to keep iterating on SWE Bench Pro rather than retire it.
RUBRIC DROPOUT: A SIMPLE WAY TO MITIGATE REWARD HACKING IN RUBRIC-AS-REWARD RL
Rubric-based RL can quietly learn to game its reward: the training judge keeps assigning higher scores even as true quality declines. A one-line fix, inspired by neural-network dropout, mitigates the problem at virtually no additional cost.
Introducing READY: What It Takes to Deploy an AI Agent
Today we're introducing READY (Reliable Enterprise Agent Deployment), a suite of industry-specific benchmarks that measure agents the way enterprises actually use them: working alongside people, inside real workflows.
Who Grades the Graders? Rethinking Verifier Design for Computer Use Agents
As Computer Use Agents (CUA) take on complex professional tasks like processing emails, creating spreadsheets, drawing 3D diagrams, and drafting financial memos, verifiers are at the heart of providing evaluation and training signals around their capabilities. Benchmarks like OSWorld 2.0 and Agents' Last Exam use these verifiers, in the form of programmatic checks, to measure whether an agent is capable of completing production-grade work.
View allAll posts