Google launches new coding model, Gemini 3.8 Flash

Смотреть на YouTube ↗  |  02 сентября 2026, 18:13  |  1:36  |  CNBC
Спикеры
Mackenzie Sigalos — Крипто-репортер/аналитик, CNBC
Google launched Gemini 3.8 Flash, its third Flash release in six weeks, positioning it as a lower-cost coding and reasoning model that approaches frontier performance. The report discusses Google's need to reset the AI narrative after a four-month share losing streak and talent departures. MacKenzie Sigalos highlights intensifying competition with Anthropic, falling token prices, and Google's high-margin pricing advantage. - Google stock is on a four-month losing streak, the longest since 2015. - Gemini 3.8 Flash is Google's third Flash release in six weeks and is pitched as its best reasoning and coding model yet. - Google says the model approaches larger frontier models on benchmarks at a fraction of the cost. - A cyber version is limited to trusted government and enterprise defenders. - Anthropic has taken coding momentum since Gemini 3 and is touting roughly 50% cheaper agentic costs. - Google is leaning into smaller Flash models that are faster to ship and cheaper. - MacKenzie notes Google has more room to compete on price because of its high-margin business.
Идеи
Mackenzie Sigalos Крипто-репортер/аналитик, CNBC 0:31
Google's AI cost edge may reset narrative.
Google is trying to regain the AI coding battleground with Gemini 3.8 Flash, its third Flash release in six weeks. The model is described as Google's best reasoning and coding model yet, approaching much larger frontier models on performance benchmarks at a fraction of the cost. Google is leaning into smaller Flash models that are faster to ship and cheaper, which may be where the market is headed. Token prices are at record lows and Google has more room to compete on price because of its high-margin business, giving it a potential edge to reset the stock's losing streak.
Далее

This CNBC video, published September 02, 2026, features Mackenzie Sigalos discussing GOOGL. 1 trade idea extracted by AI with direction and confidence scoring.

Speakers: Mackenzie Sigalos  · Tickers: GOOGL