Model ML completes finance work more efficiently with GPT-5.6 Sol

Model ML, a software company founded by cofounders Arnie and Chaz Englander, is utilizing OpenAI's GPT-5.6 Sol to power agents that automate end-to-end finance workflows. Beginning with a brief and source material, the agents handle research, analysis, and calculations to produce editable PowerPoint decks and Excel workbooks with traceable sources. Model ML evaluated GPT-5.6 Sol across multiple finance benchmarks against models such as Opus 5 and Fable 5. According to the company, GPT-5.6 Sol completed PowerPoint workflows in 100 percent of test cases, compared to 76 percent for Opus 5, and met the professional-readiness gate in 43.3 percent of cases versus 26.7 percent. Additionally, GPT-5.6 Sol used 36 percent fewer tokens per workbook than Opus 5 in Excel tasks and 21 percent fewer tokens than Fable 5 in PowerPoint workflows. These composite evaluation results prompted Model ML to expand GPT-5.6 Sol into production, including workflows previously handled by Opus 4.8. Operating within Model ML's agent harness and paired with native document tooling, GPT-5.6 Sol helps reduce the time required for deliverables, such as tearsheets that previously took an hour now taking about five minutes. The platform also enables finance professionals to process virtual data rooms containing over 100,000 rows and hundreds of files in a single pass. Following on-site sessions with OpenAI to refine agent instructions and tool loading, Model ML offers secure, interactive outputs that allow reviewers to inspect source materials and collaborate with agents directly from documents.