Google doubles down on cheaper, faster AI — but says Gemini 3.5 Pro still isn't ready (www.businessinsider.com)

Google has announced the rollout of new AI models within its Gemini series, focusing on increased efficiency and cost-effectiveness rather than raw performance. Despite the anticipation surrounding Gemini 3.5 Pro, originally expected to launch in June, it remains in testing. The new models include Gemini 3.6 Flash, touted as the "workhorse" for coding tasks and boasting a token usage reduction of up to 17% compared to its predecessor. There's also 3.5 Flash-Lite, the fastest and most cost-effective version yet, and 3.5 Flash Cyber, aimed at identifying cybersecurity vulnerabilities while remaining economically competitive. This shift in focus comes as many businesses seek to manage their AI spending amid rising token costs, creating an opportunity for Google to capture market share with its more affordable models. By emphasizing efficiency, Google is responding to a critical demand in the AI/ML community for cost-effective solutions as companies reassess their budgets in light of current economic pressures. Meanwhile, Gemini 4, the next flagship model, is under ambitious pre-training, but its release date remains uncertain. As the market evolves, Google's strategic pivot may position it favorably against competitors like Anthropic and OpenAI, who have also introduced specialized models recently.