Uploaded August 2026 | Updated September 2026, 3 weeks ago
Model ML is solving the “last-mile problem” in finance, carrying asks from research to analysis to a finished PowerPoint deck or Excel workbook with traceable sources.
When Model ML tested GPT‑5.6 Sol in its Composite, the results “blew every metric out the water,” says Co-founder and CEO Chaz Englander:
⚡ 21% fewer tokens per PowerPoint deck than Fable 5
✅ A 16.6 percentage-point lead over Opus 5 on decks ready for substantive review
📊 36% fewer tokens per Excel workbook than Opus 5
Watch Englander walk through Model ML’s evaluation process, then read the full story about how they use the model in production here: openai.com/index/model-ml
Model ML is solving the “last-mile problem” in finance, carrying asks from research to analysis to a finished PowerPoint deck or Excel workbook with traceable sources.
When Model ML tested GPT‑5.6 Sol in its Composite, the results “blew every metric out the water,” says Co-founder and CEO Chaz Englander:
⚡ 21% fewer tokens per PowerPoint deck than Fable 5
✅ A 16.6 percentage-point lead over Opus 5 on decks ready for substantive review
📊 36% fewer tokens per Excel workbook than Opus 5
Watch Englander walk through Model ML’s evaluation process, then read the full story about how they use the model in production here: openai.com/index/model-ml
