Harmonic Launches AI Chatbot App - Robinhood CEO's Startup

Harmonic Launches Aristotle AI Chatbot in Beta

Harmonic, a new AI company established with the involvement of Robinhood CEO Vlad Tenev, has announced the beta release of its Aristotle AI model. The model is accessible through a new chatbot application for both iOS and Android platforms.

Expanding Access to Aristotle

The company’s primary goal with this launch is to make Aristotle more widely available. Harmonic asserts that its AI delivers “hallucination-free” responses specifically for questions requiring mathematical reasoning. This is a significant claim, considering the known reliability issues present in current AI models.

Focus on Mathematical Superintelligence

Harmonic is dedicated to the creation of what it terms “mathematical superintelligence,” or MSI. The startup envisions a future where Aristotle can assist users across various disciplines heavily reliant on mathematical principles, including areas like physics, statistics, and computer science.

Formal Verification of Outputs

“Aristotle represents the first product available to the public that performs reasoning and formally verifies its results,” explained Tudor Achim, CEO and co-founder of Harmonic, in a recent TechCrunch interview. “Within the quantitative domains supported by Aristotle, we confidently guarantee the absence of hallucinations.”

Future Plans for Aristotle

Harmonic intends to eventually offer an API, enabling enterprises to integrate Aristotle into their own systems. Additionally, a web application is planned for broader consumer access.

IMO Performance

Aristotle demonstrated gold medal-level performance on the 2025 International Math Olympiad (IMO) through a formal testing process. This involved translating the problems into a format readable by machines. Both Google and OpenAI have also developed AI models achieving gold medal status on this year’s IMO, but their testing was conducted informally using natural language.

Benchmark Data Availability

Currently, Harmonic is not disclosing additional benchmark results for Aristotle.

Recent Funding and Valuation

The beta launch follows a recent Series B funding round where Harmonic secured $100 million, led by Kleiner Perkins, resulting in an $875 million valuation. Achim believes this valuation reflects investor confidence in Harmonic’s rapid progress towards achieving MSI and the ambitious scope of the startup.

Industry-Wide Interest in AI and Math

Numerous leading technology companies are actively developing AI models capable of solving mathematical problems. The ability to perform mathematical calculations is valuable in itself, but mathematics is also considered a uniquely verifiable field that necessitates strong reasoning abilities. Developing these capabilities could have broader applications across other domains.

How Harmonic Achieves Accuracy

Harmonic’s high level of accuracy is achieved by having Aristotle generate responses using the open-source programming language Lean. Before presenting an answer to users, the model rigorously verifies its correctness through an algorithmic process independent of AI. The CEO of Harmonic points out that similar verification technology is employed in critical fields like medical devices and aviation.

The Challenge of Hallucination-Free AI

Achieving hallucination-free performance from an AI model, even within a specific domain, is a remarkably challenging undertaking. Research indicates that even the most advanced AI models exhibit hallucinations frequently, and this issue doesn't appear to be diminishing. In fact, OpenAI’s newest reasoning models demonstrate a higher rate of hallucination compared to their predecessors.