AitopkitIvo, the contract intelligence platform for enterprise legal teams, on October 1 became the first legal AI company to release an open-source model post-trained for long-horizon contract work. Announced at Ivo Inscribe, the company's first user conference in San Francisco, Ivo Sage gives attorneys, researchers and developers a model they can download, run and build on for free — fine-tuning it on their own datasets and adjusting its rules and safeguards for specific use cases. "The next leap for AI in contract and legal work won't come from bigger models," said co-founder and CEO Min-Kyu Jung. "It will come from giving models the context of the work."
Sage was built in partnership with River AI by post-training DeepSeek V4 Flash on long-horizon contract work, using public data plus synthetic data generated by real attorneys, with River providing the training infrastructure. After reinforcement learning, the model went from scoring 70% to 91% of the pass criteria on the Legal Agent Benchmark (LAB) Contracts — reaching comparable quality to much larger frontier models with higher token efficiency, at a fraction of the cost.
Ivo also previewed a new contract-review benchmark built with research partner micro1 — one that grades judgment, not just edits: which issues to prioritize, when to show restraint, how to adapt to a deal, when to escalate to an attorney, and how to follow a playbook. Early findings are sobering for AI cheerleaders: models act like pushovers, get counter-or-reject calls right only 46% of the time, and meet just 23% of the escalation criteria set by expert attorneys. The company also announced general availability of Ivo Collaborate, an end-to-end contract intelligence platform that manages a contract's full life from intake to signature. For a vertical that has kept its models tightly closed, an open-weight release is a real shift.