Scale AI
A crowdsourced data annotation platform that provides AI training data labeling services for technology companies.
Core Facts
Timeline (last 90 days)
Scale AI's AgentEnv framework can now run on Vercel Sandbox.
AgentEnv is an agent evaluation framework launched by Scale AI that puts agents through real tasks and scores them based on performance.
The key early funding for both Scale AI and Cognition AI came from Founders Fund.
Alexander Wang dropped out of MIT to found Scale AI, focusing on AI training data annotation.
Humanity's Last Exam is an advanced reasoning benchmark jointly launched by Scale AI and the Center for AI Safety, released in early 2025, containing about 3,000 expert-level questions.
Data cleaning platforms such as Scale AI exist, but they serve large enterprise budgets, and independent researchers and individual developers are basically excluded.
A year ago Zuckerberg spent billions to buy Scale AI, put twenty-something Alexandr Wang in the position of head of Meta Superintelligence, and ousted Yann LeCun because of his fierce criticism of LLMs.
Data brokers (such as Bright Data and parts of Scale AI's business) focus on data collection, cleaning, and on-demand supply.
The crowdsourced micro-task model of Amazon Mechanical Turk is primarily used for purely digital work such as labeling and classification.
Meta acquired Scale AI and its founder Alexander Wang for approximately $15 billion, and also recruited Nat Friedman.
17 more timeline events
All Facts (20)
Professional annotation companies such as Scale AI and Surge AI provide higher-quality structured annotation services.
75%UnverifiedDatabricks reached a valuation of $43 billion in 2023.
70%UnverifiedCognition was co-founded by Scott Wu, a former Scale AI executive.
60%UnverifiedAlexandr Wang is the founder and CEO of AI data annotation company Scale AI, who started the company at age 19, with the company's valuation exceeding $13 billion.
60%UnverifiedAgentEnv is an agent evaluation framework launched by Scale AI that puts agents through real tasks and scores them based on performance.
50%UnverifiedScale AI's AgentEnv framework can now run on Vercel Sandbox.
50%UnverifiedAlexander Wang dropped out of MIT to found Scale AI, focusing on AI training data annotation.
50%UnverifiedThe key early funding for both Scale AI and Cognition AI came from Founders Fund.
50%UnverifiedHumanity's Last Exam is an advanced reasoning benchmark jointly launched by Scale AI and the Center for AI Safety, released in early 2025, containing about 3,000 expert-level questions.
50%UnverifiedData cleaning platforms such as Scale AI exist, but they serve large enterprise budgets, and independent researchers and individual developers are basically excluded.
50%UnverifiedData brokers (such as Bright Data and parts of Scale AI's business) focus on data collection, cleaning, and on-demand supply.
50%UnverifiedThe crowdsourced micro-task model of Amazon Mechanical Turk is primarily used for purely digital work such as labeling and classification.
50%UnverifiedMeta acquired Scale AI and its founder Alexander Wang for approximately $15 billion, and also recruited Nat Friedman.
50%UnverifiedThe article states that the HELM evaluation framework was released by Scale AI.
50%UnverifiedHLE (Humanity's Last Exam) is an extremely high-difficulty comprehensive knowledge test constructed by Scale AI in collaboration with numerous domain experts.
50%UnverifiedCloud-based SaaS annotation platforms such as Labelbox and Scale AI are powerful but require paid subscriptions and raise data-upload concerns.
50%UnverifiedScale AI became a data supplier to leading AI laboratories through standardized data annotation services, serving as a precedent that ALIA can reference
50%UnverifiedScale AI's SEAL leaderboard and LMSYS's Chatbot Arena adopt closed evaluation strategies; Chatbot Arena evaluates models through an Elo scoring mechanism based on blind-test votes from real users.
50%UnverifiedHLE (Humanity's Last Exam) is an extremely high-difficulty comprehensive reasoning benchmark jointly launched by Scale AI and multiple academic experts, covering dozens of disciplines including mathematics, physics, biology, philosophy, and law.
50%UnverifiedThe Cognition team was co-founded by Scott Wu, a former Scale AI executive.
50%