Ivo, a contract intelligence platform serving major global enterprises, announced it has become the first legal AI company to release an open-source model post-trained for long-horizon contract work. The announcement was made at Ivo Inscribe, the company’s first user conference held in San Francisco on October 1, 2026.
The newly named Ivo Sage model gives developers, researchers and legal teams access to AI technology they can test, train and build on. By open-sourcing its model, the company is enabling legal practitioners, AI researchers and developers to fine-tune it on their own datasets, adjust its rules and safeguards for specific use cases, and iterate quickly on new ideas or approaches.
“The next leap for AI in contract and legal work won’t come from bigger models,” said Min-Kyu Jung, co-founder and CEO of Ivo. “It will come from giving models the context of the work: the documents, the playbooks and the way a legal team operates.” Jung emphasized that contextual understanding represents the hard, unsolved part of legal AI where the company is focused.
The Ivo Sage model was built in partnership with River AI by post-training DeepSeek V4 Flash on long-horizon contract work using both public and synthetic data generated by real attorneys, with River providing valuable technical and training infrastructure support. After reinforcement learning, the model improved from scoring seventy percent of the pass criteria of the Legal Agent Benchmark (LAB) Contracts to ninety-one percent, reaching comparable quality to much larger frontier models while maintaining higher token efficiency at lower cost.
Related Post: Mishandling gender identity data costs life sciences millions
A New Contract Review Benchmark
While other legal benchmarks grade only the edits a model makes, Ivo’s new benchmark builds better judgment and escalation reasoning into the review process. Contract review is a hard test for AI because most judgment happens before any edit is proposed. A good reviewer must know when to propose a change, when to leave acceptable language alone and when a decision needs to be escalated for an attorney’s judgment call.
At Inscribe, Ivo previewed the Ivo-micro1 Contract Bench configured to measure attorney-like distinctions across five dimensions: which issues to prioritize, when to show restraint, how to adapt to the specific deal, when to escalate, and how it follows your playbook. Ivo’s research found that leading models handle certain contract decisions well and miss others entirely, especially when to push back and when to escalate.
Models are pushovers when facing counterparty changes. They concede easily and push back poorly. When the right response is to accept or leave language alone, models get it right 82 percent of the time. When it’s to counter or reject, they get it right only 46 percent of the time, adding a needed new point of their own just 23 percent of the time.
They also struggle with knowing when to escalate. On average, they meet only 23 percent of the escalation criteria set by expert attorneys, despite explicit instructions and dedicated escalation tools. This indicates models make edit decisions independently rather than seeking permission from authority. GPT-6 Astra stands as an outlier at 71 percent escalation rate, but it meets only 31 percent of its edit criteria and finishes last overall.
They make edits in broad strokes instead of the precise inline changes attorneys prefer. Attorneys make 64 percent of their changes inline, averaging 92 characters. Models make only 14 to 37 percent of changes inline, rewriting whole blocks of text at 207 to 369 characters.
Read Also: Lawyers Leave Firms for AI Jobs
Every model performs worse when deals require deviation from standard playbooks. Models meet 51 percent of criteria for applying playbook standard positions, but this drops to 38 percent when the right move depends on deal-specific facts, representing a 9 to 15 point decline.
A public set of the benchmark results will be available in the coming weeks.
Expanding Contract Lifecycle Management Capabilities
Ivo also announced the general availability of Ivo Collaborate, an end-to-end contract intelligence platform designed for enterprise legal teams. The platform manages the entire contract lifecycle, from initial intake through signature and beyond, enabling automated workflow routing based on predefined rules. Routine agreements are processed according to legal-established protocols while flagging complex matters requiring human review.
Enterprise-Wide Contract Intelligence Vision
Ivo’s technology helps organizations instantly surface contractual risks, streamline agreement execution and enhance overall business performance. Founded in New Zealand and headquartered in San Francisco, Ivo enables organizations to surface risk instantly, accelerate negotiation cycles, and remove bottlenecks across the contract lifecycle. Customers include leading global enterprises and high-growth companies across multiple industries.
