Skip to main content

Prentis

Category: AI Agents

An AI research lab building computer-use models that perceive screens and operate software to automate enterprise office workflows. Prentis was founded in 2026. The company is led by Ritankar Das. Based in San Francisco Bay Area, California, United States. Team size: 11-50. Latest round: Undisclosed.

Founded
2026
Headquarters
San Francisco Bay Area, California, United States
Team size
11-50

Value proposition

AI agents that can control computers to automate routine enterprise workflows at roughly 10x lower cost per task than frontier APIs, using a smaller, more economical model (Hive-32B) that outperforms larger models on computer-use benchmarks

Products and solutions

Hive-32B model (computer-use AI model), Enterprise AI agents for automating office workflows including insurance claims processing, customs duty refund exceptions, and other back-office drudgery across retail, healthcare, and manufacturing

Unique value

10x lower cost per task than frontier APIs; Hive-32B model outperforms GPT-5.4 and Claude Opus 4.6 on WindowsAgentArena and ScreenSpot-v2 benchmarks; trains models inside actual workflows rather than on public web data

Target customer

Large enterprises in retail, healthcare, and manufacturing with complex back-office workflows (insurance claims, customs duty refunds, document processing)

Industries served

Healthcare, Manufacturing, Retail, Customs/Trade

Technology advantage

Hive-32B model — a smaller, cheaper model (32B parameters) that claims superior performance on computer-use benchmarks (WindowsAgentArena, ScreenSpot-v2) vs. GPT-5.4 and Claude Opus 4.6; continuous learning loop of perceiving interfaces, acting, and learning from outcomes; trains on proprietary operational data rather than public web data

How they differentiate

Runs a much smaller, cheaper model (32B parameters) vs. frontier models, claiming ~10x lower cost per task; trains on actual enterprise workflow data rather than public web data; performance-dependent pricing model (20% of savings realized)

Main competitors

OpenAI (GPT-5.4 / computer-use agents), Anthropic (Claude Opus 4.6 / computer-use agents), Thinking Machines Lab (Mira Murati's startup)

Notable customers

Healthcare management services organization, Manufacturer, Goods and clothing manufacturers (signed contracts worth up to $50M total)

Major milestones

Launched in April 2026, Signed contracts worth up to $50M with enterprise customers, Hive-32B model outperforms GPT-5.4 and Claude Opus 4.6 on computer-use benchmarks, In talks to raise $100M at $1B valuation (as of July 2026), Hired 25+ employees from OpenAI, Google DeepMind, Meta, Tencent, and Alibaba

Growth metrics

Projected ~$75M annualized run rate by Q3 2026 (based on 20% fee of savings realized, performance-dependent, not recognized revenue); Contracts worth up to $50M signed

Market positioning

Early-stage AI research lab positioning as a cost-efficient alternative to frontier model providers for enterprise computer-use automation, backed by high-profile tech founders

Geographic focus

United States (San Francisco Bay Area)

About Ritankar Das

Founder of Titan (holding company building AI startups); Gates Cambridge Scholar; dropped out of AI PhD at Cambridge; UC Berkeley's youngest University Medalist in over a century, graduating at 18 with double major in bioengineering and chemical biology; MSc in Biomedical Engineering at Oxford

Latest news about Prentis

More AI Agents companies

Official website: