Model ML raises $250M Series B at $1.5B valuation for banking workflow agents
London-based Model ML said on November 17 that it raised $250 million in a Series B round at a $1.5 billion valuation. General Atlantic and QED Investors co-led the round, with QED already an investor. TechCrunch reported that the startup builds agents that take on work usually given to junior bankers and consultants, such as financial models, pitch decks and due diligence reports. The raise is more than three times the $75 million Series A announced about a year earlier.
The agents connect to a client's systems, data and tools, including Excel and PowerPoint, and revise financial models and decks when deal facts change. The platform runs inside a firm's own infrastructure and routes tasks across frontier and fine-tuned open-source models. Model ML says its open-source stack completes applicable work at about 2% of frontier model cost, roughly five times faster, with comparable accuracy on those tasks. The company says customer data is never used for training and holds ISO/IEC 27001:2022 certification plus a SOC 2 Type II audit.
Chaz Englander is CEO, and he co-founded the three-year-old company with his brother Arnie, according to FinSMEs. Their earlier business Fat Llama, a peer-to-peer rental marketplace, was sold to Hygglo in 2022 for $41.5 million. New capital will fund research and engineering in London and New York and larger customer deployments. The company also reports that annual recurring revenue grew 12x year on year in Q3 2026 and daily active users rose 26x, but it has not disclosed revenue totals.
Why it matters
For founders building vertical agents, the round suggests a professional-services product can reach a $1.5 billion valuation about three years after founding. The more useful lesson is the cost claim. If open-weight models deliver comparable results at a fraction of frontier pricing, cost per completed task becomes a pricing and margin lever, especially for regulated buyers. Routing work across several models also reduces dependence on any single vendor's pricing changes.
For sellers, banks and consultancies are testing whether agents can run inside their own infrastructure with audit-grade security, so certifications and no-training commitments are part of the pitch. For investors, a more than threefold step-up within a year is a strong signal, but it has not been paired with disclosed revenue. The 2% cost and five-times-faster figures are company claims that buyers will want to test on their own workloads.
Key points
- 1Total funding reaches $350 million, up from a $75 million Series A announced about a year earlier.
- 2Model ML says its open-source stack runs at about 2% of frontier model cost and roughly five times faster.
- 3The company reports annual recurring revenue grew 12x year on year in Q3 2026, with daily active users up 26x.
- 4Fat Llama, co-founded by the Englander brothers, sold to Hygglo in 2022 for $41.5 million.
- 5