I'm thrilled to introduce Fintool v4, the most powerful agent for equity research and investment. On the hardest benchmark for equity analysts, Finance Agent Benchmark, it scores 63% better than the second best AI while being 4x faster and 10x cheaper.
In this post, I want to cover:
The Finance Agent Benchmark, developed by vals.ai in collaboration with Stanford researchers, a Global Systemically Important Bank, and industry experts, addresses this gap by testing AI models on real-world financial research tasks that require complex analysis using recent SEC filings.
Here are some examples of the types of questions in the benchmark:
For financial institutions running hundreds or thousands of queries per day, this cost efficiency is transformative.
A hedge fund analyst performing 50 research queries daily would spend:
Because Fintool is so fast and cheap, it can run hundreds of queries in parallel to create an industry or a company research primer.
View example →
View example →
We knew we were onto something when beta users kept telling us that Fintool answers were significantly better than every DeepResearch product out there but also 10x faster. That speed advantage is game changing because it lets analysts iterate rapidly and ask dozens of follow-up questions in the time it would take other tools to answer just one.
Edouard GodfreyChief Technology Officer at Fintool
Compared to human analysts, the difference is even more dramatic. A task that takes a junior analyst 16.8 minutes and costs $25.66 is completed by Fintool v4 in 40 seconds for $0.14.
That's 25x faster and 183x cheaper, with 90% accuracy on expert-level tasks.
Where did Fintool fail? Some are small mistakes that are easy to fix. For example, for "What are Netflix's (NASDAQ: NFLX) Total Projected Material Cash Requirements for 2025?" we used the next twelve months instead of the calendar year. The second one is more tricky but still fixable: for "What would be the impact to net income in dollars and percent if all debt for Boeing in 2024 were refinanced at 3% higher interest rates than current?" Fintool v4 did everything right but forgot to apply a tax shield.
We are confident we can reach 100%. Beyond the accuracy gains, the unit economics are what make this truly transformative. At 183x cheaper and 25x faster than a junior analyst, teams using Fintool can expand their coverage universe dramatically, ramp up on new names in minutes instead of days, and make better-informed investment decisions with unprecedented speed.
Fintool v4 now integrates consensus estimates from S&P Global, covering earnings forecasts, revenue projections, and price targets from hundreds of sell-side analysts. This allows you to instantly compare actual results against Street expectations and identify estimate revisions in real-time.
Fintool now searches the internet more broadly, but still verifies information accuracy and considers only trusted sources. This is particularly useful for tracking M&A announcements, management changes, regulatory developments, or industry trends that have not yet appeared in formal disclosures.
Fintool v4 accesses comprehensive market data including historical prices, trading volumes, corporate actions, and index constituents. This enables technical analysis, relative performance comparisons, and valuation multiples calculations using institutional-grade data.
I'm excited about the new use cases this unlocks. Teams can now automate repetitive analyst workflows: upload an earnings template and Fintool will generate it for the most recent quarter. You're not limited to one company either: generate detailed industry briefs and competitive analyses at coverage scale. Set up intelligent alerts like "On release of a new earnings call, notify me how key themes are trending over the last four calls" and let Fintool monitor developments across your entire portfolio.
Fintool v4 is able to do the job of a mid-level equity analyst. The next version we're working on, Fintool v5, will blow past senior equity research and take a shot at the job of a PM: finding relevant investment opportunities, synthesizing cross-sector insights, and surfacing alpha-generating ideas before they hit consensus.