HomeNews
Share

Alphabet's stock is rising. The company has launched a new AI model for programming

Yana Zakomoldina

Yana Zakomoldina

Reporter
Shares of Alphabet, Googles parent company, rose more than 4% in premarket trading. Photo: Ken Wolter/Shutterstock

Shares of Alphabet, Google's parent company, rose more than 4% in premarket trading. Photo: Ken Wolter/Shutterstock

Shares of Alphabet, Google’s parent company, rose more than 4% in premarket trading on November 19. Among the drivers of growth, in addition to the successful IPO of AI developer Anthropic—in which Google has invested— was the release of the new AI model Gemini 4 Neon, as noted by The Wall Street Journal. According to a post on the company’s blog, the neural network in question is primarily designed for programming, data analysis, scientific work, and office tasks. Within the Gemini 4 family, Neon occupies a position between Carbon—designed for the most complex and time-consuming projects—and Hydrogen—intended for quickly performing everyday tasks, according to Google’s blog.

What is known about the new model

Google has unveiled Gemini 4 Neon—an AI model designed for programming, analytics, and scientific work. The neural network processes audio, video, graphics, PDF files, and spreadsheets. According to Glean, one of the first companies to test Neon, the model is on par with the more powerful Carbon in terms of response quality, but it runs “three times faster and costs 60% of Carbon’s price.”

One of Neon’s unique features is that it has a separate version designed for cybersecurity. According to Google, it writes “more secure code” than the more powerful Carbon. In addition, it can detect vulnerabilities in code, verify whether they can actually be exploited, and fix them. According to the company’s press release, only verified security experts—participants in the Fairwind program—will gain access to this version in the coming days.

According to the independent platform Artificial Analysis, cited by Google, Neon ranked first among AI models for code generation and was among the top three in the overall “intelligence” ranking (as part of this study, AI models undergo a series of tests on reasoning, mathematics, programming, and general knowledge, and the results are combined into a single score, referred to as the Intelligence Index). Google itself also stated that, in internal tests, the model demonstrated “the lowest level of fabrication among all of the company’s models to date.”

Google has already begun rolling out access to Gemini 4 Neon for developers, enterprise customers, and paid subscribers.

What People Are Saying in the Market

The tech publication SiliconANGLE views Gemini 4 Neon as Google’s response to competition from OpenAI and Anthropic, which released powerful new models this summer. At the same time, the publication highlights the price: Neon is more expensive than OpenAI’s GPT-6.2 Sol model, though it is cheaper than Anthropic’s Claude Opus 5.5. Since models in this class are comparable in capabilities, many companies may decide that Neon’s advantages do not justify the higher price, SiliconANGLE writes.

For their part, analysts at the investment firm Golden Door Asset emphasize the new model’s cost-effectiveness. They note that the Neon even slightly outperforms the Carbon in the Artificial Analysis rankings, while costing 25% less than the flagship model’s standard pricing. According to the analysts, Neon effectively replaces the previously released Argon model, leaving Google with a clear lineup of three models for tasks of varying complexity and budgets.

This article was AI-translated and verified by a human editor

Share

Trending

Stock Screener
Buy
Sell
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
‌
Small Caps
Investment and Finance News