Computer-Use AI Agents

AI Lab Prentis Launches Its Hive-32b Model

Prentis, an AI lab co-founded by Reid Hoffman, Mark Pincus and CEO Ritankar Das, launched in April to build computer-use AI agents, featuring Hive-32B — a smaller model trained to operate desktop apps and replicate routine office workflows. The company said Hive-32B outperformed larger rivals on internal benchmarks that measure end-to-end task completion and on-screen control detection.

The startup is positioning Hive-32B as a cost-efficient alternative to frontier APIs, claiming roughly 10 times lower cost per task and already signing contracts reportedly worth up to $50 million with customers in healthcare and manufacturing. Its approach centers on task-specific agents that automate claims processing, customs exceptions and other document-driven workflows.

For users, Prentis promises hands-free completion of repetitive computer tasks, reducing time spent on admin work and enabling staff to focus on judgment-heavy work. If its performance and commercial deals hold, the launch underscores a broader trend toward agentizing everyday office processes rather than replacing core developer workflows.

Image Credit: Prentis

Computer-use Agents
AI models that operate desktop applications directly are creating new openings for automating routine office workflows without requiring deep software integrations.
Task-specific AI
Smaller models optimized for narrow business processes are reshaping enterprise AI economics by lowering per-task costs while maintaining practical performance.
Document Workflow Automation
Claims, customs and compliance-heavy processes are becoming fertile ground for AI agents that can interpret screens, handle exceptions and complete repetitive administrative steps.

Industries Being Reshaped

Healthcare
Administrative workloads such as claims processing and patient documentation present major potential for AI agents that reduce manual handling and accelerate back-office operations.
Manufacturing
Complex supply chain paperwork and customs exceptions create opportunities for computer-use AI to streamline document-driven coordination across distributed operations.
Enterprise Software
Agent-based automation is expanding the role of workplace platforms by embedding AI systems that perform tasks across existing applications rather than replacing them.
SCORE
3.7 out of 10
GENDER
50% Men50% Women
MARKETTop markets: North America, Europe, Asia
GENERATION
  • Gen Z
  • Gen Alpha
  • Millennial (primary audience)
  • Gen X (primary audience)
POPULARITY
Popularity 11%
Activity 0%
Freshness 100%