AI can now pass the hardest level of the CFA exam in a matter of minutes

New research from NYU Stern and GoodFin reveals that advanced large language models, including o4-mini, Gemini 2.5 Pro, and Claude Opus, can now pass CFA Level III mock exams in minutes using 'chain-of-thought prompting,' overcoming previous AI struggles with essay questions. This demonstrates AI's rapidly evolving capability for complex financial analytical reasoning and portfolio management, signaling a potential transformation in the industry, though human judgment for context and intent remains a critical differentiator.

Analysis

New research indicates that frontier large language models, including o4-mini, Gemini 2.5 Pro, and Claude Opus, can now successfully pass the Chartered Financial Analyst (CFA) Level III mock examination, a task previously insurmountable for AI due to its focus on complex essay questions related to portfolio management and wealth planning. This breakthrough, achieved in minutes using 'chain-of-thought prompting,' demonstrates a significant leap in AI's capacity for the specialized, high-stakes analytical reasoning required in professional finance, contrasting sharply with the 1,000 hours of human study typically needed. While this rapid technological evolution signals a fundamental transformation in the financial industry, expert commentary from GoodFin's CEO suggests AI is unlikely to completely replace human analysts. The core human advantage remains in assessing client-specific context, intent, and non-verbal cues, which are critical for bespoke wealth management and are currently beyond the scope of machine assessment. The development positions AI as a powerful tool for augmenting, rather than replacing, financial professionals by automating complex analytical tasks and drastically improving efficiency.

Key Decisions for Investors

• Investors should increase exposure to the AI and Fintech sectors, particularly firms developing or effectively integrating advanced reasoning models for financial analysis, as these are positioned to capture value from the imminent industry transformation.
• It is prudent to re-evaluate holdings in traditional asset and wealth management firms, scrutinizing their AI adoption strategies, as companies failing to leverage these new efficiencies may face competitive disadvantages and margin compression.
• Monitor for emerging opportunities in companies that provide the technological infrastructure for these advanced AI models, as their platforms are now proven to handle sophisticated, domain-specific reasoning, suggesting broad applicability and a strong growth trajectory.