Google’s Cheaper Gemini AI Push Collides With Embarrassing Gemini 3.5 Pro Delay

• June promise, July delay
• July 21: Cheaper Gemini models arrive instead
• Frontier gap and market skepticism
• July 22: Rivals pounce
• Strategic bet: efficiency over bragging rights

Google’s Cheaper Gemini AI Push Collides With Embarrassing Gemini 3.5 Pro Delay
Google is racing to make AI cheaper and more efficient just as rivals try to paint its most advanced model as missing in action, turning a product roadmap into a public perception battle.

§June promise, July delay

At Google I/O in May, the company told developers that Gemini 3.5 Pro was already being used internally and was expected to roll out in June. By mid‑July, however, the model still had not shipped, and reports noted Google had been “working on beefing up the model’s skills, particularly in coding,” without explaining the holdup.

§July 21: Cheaper Gemini models arrive instead

On July 21, Google formally launched three new Gemini models — 3.6 Flash, 3.5 Flash‑Lite, and 3.5 Flash Cyber — all pitched as faster and more cost‑efficient options. The AI deployment race, one analysis noted, has shifted “from benchmark bragging rights to who can provide the best model at the lowest price.”

Gemini 3.6 Flash replaces 3.5 Flash as the “workhorse,” improving coding and multimodal performance while cutting token usage by up to 17% versus its predecessor. Google describes 3.5 Flash‑Lite as its “fastest and most cost‑effective” 3.5‑series model, targeting high‑volume agents and document processing. Executives amplified the message on X, saying the launches are “all about better performance, lower latency, and a smaller bill,” highlighting that 3.6 Flash can cut token usage by up to 65% on complex coding while 3.5 Flash‑Lite reaches 350 output tokens per second.

In security, Google introduced Gemini 3.5 Flash Cyber, a “cost‑efficient and highly capable alternative” to larger, expensive models like Anthropic’s Mythos, integrated into its CodeMender agent to rapidly scan code for vulnerabilities. DeepMind framed the model as a lightweight system that lets defenders call it “multiple times at high speed and low cost” to explore more code paths and patch flaws.

§Frontier gap and market skepticism

Yet the same day’s coverage emphasized what was missing: “Google releases three new Gemini models — but no 3.5 Pro,” one headline noted, underscoring fresh questions about its AI strategy. Business press framed the line‑up as evidence that Google is doubling down on cheaper, faster AI because “Gemini 3.5 Pro … is still in testing,” even as Gemini 4 is only in pre‑training.

Alphabet’s shares dipped around the announcement window, and Axios reported that the absence of 3.5 Pro — rumored to be months behind schedule — comes as Google tries to retain key DeepMind researchers amid a talent war.

§July 22: Rivals pounce

On July 22, competitors seized on the opening. A Meta executive mocked Google with the post “gemini who?” after Meta’s Spark model topped a leaderboard over a Google system. An OpenAI engineer, responding to news that Gemini 4 pre‑training had begun, quipped, “Hope it finishes one day too!” in a jab at the slipping timelines.

Analyst Josh Beck argued the delay has “shifted perception from leading edge to trailing edge” for Google’s frontier AI, even as he cautioned it is “too early to count anyone out” in a fast‑moving field. At the same time, some users praised Google’s focus on efficiency, saying models like Gemini 3.5 Flash are already their “daily driver” for high‑value tasks such as document extraction.

§Strategic bet: efficiency over bragging rights

Across the coverage, a common thread is that enterprises are hitting budget limits on token‑hungry frontier models and hunting for a “sweet spot” of price and performance. Google’s CEO has warned that companies are “blowing through their annual token budgets,” arguing a mix of lighter Flash models and frontier systems could save “a lot of money.”

Whether this efficiency‑first portfolio offsets the reputational hit from a repeatedly delayed Gemini 3.5 Pro remains unresolved. For now, Google’s cheaper models are shipping, its flagship is not, and rivals are determined to keep that contrast in the spotlight.

Continue reading https://foxvector.com/stories/019f8c7e-a103-056a-73c9-381fb64eab6f