Back to News
Market Impact: 0.1

These 7 AI models just overtook ChatGPT in a new study — and the list may surprise you

Artificial IntelligenceTechnology & InnovationProduct LaunchesAntitrust & CompetitionAnalyst Insights
These 7 AI models just overtook ChatGPT in a new study — and the list may surprise you

Prolific's Humaine benchmark, published in September, ranked generative AI chat models by human-centric evaluation and placed Google Gemini 2.5 Pro first while listing OpenAI's ChatGPT-4.1 in 8th position. The study—designed to reduce sample bias with automated quality monitoring—underscores intensifying competition from Google, DeepSeek, xAI (Grok) and Mistral and highlights potential product-performance pressure on OpenAI even as newer models from competitors were not yet available at the time of testing.

Analysis

Market structure: Google (Alphabet/GOOGL) is the primary beneficiary of a perceived product-quality lead, increasing its leverage to capture incremental AI search/assistant monetization while putting pressure on OpenAI-linked Microsoft (MSFT) to defend product parity. Compute and cloud providers (NVDA, AMZN, GOOGL Cloud, MSFT Azure) gain pricing power as demand for higher-quality LLM inference rises; expect GPU/accelerator scarcity to keep NVDA revenue growth above industry average for 2–4 quarters. Cross-asset: widening equity dispersion will raise tech single-name implied vols and tighten IG tech credit spreads; higher data-center power consumption supports select energy names (e.g., EXC) and raises industrial capex for TSMC/ASML suppliers over 12–24 months. Risk assessment: Tail risks include rapid regulatory intervention (EU/US antitrust or AI safety rules) that could force product constraints within 3–18 months, and a sudden NVDA supply shock or Chinese competitor release that compresses margins. Immediate (days): headline-driven volatility; short-term (weeks–months): product rollouts and guidance revisions; long-term (quarters–years): monetization of assistants into ads/subscriptions. Hidden dependencies: model leadership is fungible without superior data or distribution — monetization depends on ad/enterprise integration and enterprise cloud contracts. Trade implications: Favor selective longs in GOOGL and NVDA to capture product + compute tailwinds, and express relative weakness in MSFT via defined-risk put spreads rather than outright shorts to account for Azure diversification. Implement call-spread exposure to GOOGL over 3–9 months to capture upside while financing premium via modest covered-call hedges; rotate 2–4% portfolio weight from high-multiple pure-play AI names into semiconductors and cloud infrastructure. Entry/exit: scale in on 5–10% pullbacks, set profit-taking at 20–30% rallies or on guidance misses >5%. Contrarian angles: The market may underprice MSFT’s durable Azure and enterprise lock-in — a persistent OpenAI performance lag does not equal MSFT loss of cloud revenue; conversely, Google’s lead may be overestimated if monetization lags by >6–12 months. Historical parallels to search/browser battles show product wins can take multiple years to translate to ad revenue, so multiples could compress if expectations run ahead of monetization. Unintended consequence: aggressive product positioning could accelerate regulatory scrutiny, creating episodic selloffs that present tactical entry points.