The world's most accurate AI for private markets data — benchmarked, traceable, and proven against frontier models.
Purpose-built for private markets. Benchmarked against the models that weren't.
henonMAX 2.1 scored 36 percentage points higher than Claude Code (Opus 4.8) on the Private Markets Agent Benchmark — 60 tasks of real private equity and credit workflows. Accuracy ties on every self-contained task, and diverges exclusively on tasks requiring stored firm policy, cross-entity conventions, or historical state.
Raw vision breaks on dense fund tables. henonMAX's dual-AI RAG pipeline: 35/36 exact. Opus vision: 7/36.
A file-only agent re-pays per close. By month 12, Claude Code cost $6.40 at 20% accuracy. henonMAX: $0.12 at 100%.
Allocation keys, fiscal calendars, proprietary add-back rules — none of it lives in a file. A file-only agent scores 0% on every task that depends on stored convention.
Every value pinned to its row, column, and bounding box on the source PDF. Vision scored 0/5 on cell tracing. henonMAX scored 5/5.
Every department in a private markets firm faces a version of the same infrastructure gap. The M-Series was built to close it.
Stop spending weeks reconciling data across administrators, banks, and portfolio companies before any analysis can begin. M1-Monitor gives you a live, verified view of your entire portfolio — always current, always accurate.
Waterfall calculations, portfolio models, and scenario analyses that live in fragile spreadsheets maintained by one or two people — that ends here. M2-Model gives every builder a zero-error foundation for the models that matter most.
LP reports stitched together from multiple sources. Last-mile errors discovered after distribution. Compliance summaries that take weeks. M3-Measure closes the gap between your data and the reports that depend on it.