Arcee AI spent half its venture capital to build an open reasoning model that rivals Claude Opus in agent tasks

THE DECODER / 4/12/2026

📰 NewsSignals & Early TrendsIndustry & Market MovesModels & Research

Key Points

  • Arcee AI reportedly allocated about half of its venture capital to train Trinity-Large-Thinking, an open reasoning model with 400B parameters aimed at excelling in agent-style tasks.
  • The startup positions Trinity-Large-Thinking as a competitor to Claude Opus specifically on agent-task performance.
  • The article frames the move as a significant scaling effort, signaling continued investment in large open reasoning models rather than closed alternatives.
  • By emphasizing “open” availability and agent competence, Arcee AI is targeting developers and teams that want strong autonomous/agentic capabilities with less vendor lock-in.

Turquoise triangle logo above Trinity Large Thinking and subtitle Scaling an open source agent against a dark background

US start-up Arcee AI spent roughly half its total venture capital to train Trinity-Large-Thinking, an open reasoning model with 400 billion parameters designed to take on Claude Opus in agent tasks.

The article Arcee AI spent half its venture capital to build an open reasoning model that rivals Claude Opus in agent tasks appeared first on The Decoder.