How Google’s Gemini 3 Made Alphabet a Record Breaking Stock

Alphabet CEO Sundar Pichai launched Gemini 3 on Tuesday. (David Paul Morris/Bloomberg)

Key Points

• Alphabet launched its Gemini 3 model, which analysts view positively, with its shares rising 1.6% in premarket trading.
• Gemini 3 outperforms OpenAI’s GPT 5.1 and other rivals in general intelligence and coding tasks, according to Artificial Analysis.
• Gemini has over 650 million monthly active users, an increase from 450 million in July, while ChatGPT has 800 million weekly active users.

Alphabet stock rose on Wednesday amid optimism about Google’s new Gemini artificial intelligence model.

Alphabet launched its Gemini 3 model on Tuesday, claiming it is its most advanced AI yet. Wall Street and technology analysts were generally positive about its performance versus ChatGPT.

Alphabet shares closed up 3% at $292.82 on Wednesday, a new record closing high, leaving it close to taking Microsoft ‘s place as the third largest U.S. company by market valuation.

Alphabet was a Barron’s stock pick coming into 2025 and the stock is up more than 70% since then.

“In the last few months, investor sentiment has shifted significantly from doubting Alphabet’s ability to win in AI to now being discussed as a “full-stack AI winner.” Gemini 3’s release should bolster investor confidence,” wrote William Blair analyst Ralph Schackart in a research note.

The new model is going up against OpenAI’s GPT-5, the latest AI which underlies its ChatGPT app. Its release in August came as something of a disappointment, with benchmarking tests suggesting GPT-5 was only mildly better than existing rivals from Google and Elon Musk’s xAI.

“Unlike GPT-5 that had mixed reactions and…was far better on vibe tests and qualitative assessments than what was reflected in its benchmark scores, Gemini 3 has outstanding benchmark scores that are also backed up by stellar performance in our own assessments and vibe tests,” wrote D.A. Davidson analyst Alexander Platt in a research note.

OpenAI last week rolled out an update for GPT-5 to what it calls ‘5.1’ which it said would bring improvements in intelligence and communication style. Gemini 3 Pro appears to outperform GPT 5.1 across general intelligence and coding-specific tasks according to data from AI-benchmarking firm Artificial Analysis, as well as rivals from xAI, Meta Platforms and China’s Alibaba and DeepSeek, although it is also more expensive to use.

Of course, benchmark scores aren’t everything in the AI race. Many users don’t use the most advanced models and are more attached to a specific app because of its ease of use or conversational style. However, Google looks to be catching up with its rival in numbers of users.

Gemini currently has over 650 million monthly active users, up from around 450 million in July. This compares with ChatGPT’s 800 million weekly active users, up from roughly 700 million in July.

“Model advancements, historically, have had a notable benefit to user growth for AI apps and token consumption on enterprise workloads,” wrote Cantor analyst Deepak Mathivanan in a research note. “We will be watching for user growth and engagement metrics on search and Gemini app to assess the positive benefits to Google’s core business from Gemini 3 in 2026.”

Apart from the Gemini 3-powered app, Google AI Ultra subscribers will have access to an experimental feature called Gemini Agent, designed to complete tasks such as organizing an inbox or completing travel bookings. The company is also launching a new development platform called Google Antigravity for AI-powered coding.

Separately, Alphabet’s Google DeepMind said Wednesday it was opening a new AI research lab in Singapore, which will collaborate with governments, academic institutions and businesses across the Asia-Pacific. The lab will aim to improve Gemini’s capabilities, and apply the latest models across Google products and for Cloud customers.

Write to Adam Clark at adam.clark@barrons.com