Google unveiled a two‑pronged expansion of its Gemini artificial‑intelligence portfolio on Tuesday, giving AI Ultra members early access to the Gemini Spark agent platform while also preparing to replace its Veo video model with the newly branded Gemini Omni Flash. The announcements signal the company’s push to make its AI tools more versatile and commercially viable.
Spark rollout for AI Ultra members
Gemini Spark, a customizable AI‑agent framework, will begin rolling out to subscribers of Google’s AI Ultra plan next week. The service, which costs $100 a month, offers the latest Gemini features, while a higher‑priced $200 tier retains expanded token limits for power users. Google says the Spark platform will eventually be available to all Gemini users, even those who do not pay for the Ultra subscription.
Internal testing has already produced notable results. An employee identified only as Doshi described using Spark daily over the past few weeks, both for work‑related and personal tasks. In preparation for Google’s I/O conference, Doshi employed a Spark agent to compile performance statistics on the Gemini 3.5 Flash model, creating a slide deck for senior executives. “It turned out beautifully,” Doshi said, adding that the process was faster and produced higher‑quality output than a manual approach.
On the personal side, Doshi built an agent to monitor developmental milestones for her newborn. The agent analyzes data, offers insights, and suggests additional metrics to track. “I’m treating my child like an AI model,” she joked, noting the tool’s practical value despite the privacy considerations of feeding personal data to a cloud‑based system.
Omni video model takes over from Veo
Google also announced that Gemini Omni Flash will succeed Veo 3 as the company’s default video‑generation model in products such as the Gemini app, YouTube, and Flow. While Veo debuted at last year’s I/O, the new Omni model is part of Google’s broader vision for a truly multimodal AI that can accept any type of input—text, images, audio, or video—and generate corresponding output. At present, the focus is on video, which explains the transition.
Google describes Omni as an “everything model” designed for future expansion beyond video. The company emphasized that the current rollout is a stepping stone toward a more comprehensive multimodal capability, though the full range of functions remains forthcoming.
These developments arrive as Google continues to refine its AI pricing strategy. The $100 Ultra tier, while still considered high for many, reflects a $50 reduction from the previous $150 offering. The $200 tier, retained for users needing larger token quotas, also saw a $50 price cut.
Industry observers note that the combination of a more accessible Spark platform and a versatile video model could broaden the appeal of Google’s AI ecosystem, especially as competitors roll out comparable features. Whether the new pricing will attract a wider user base remains to be seen, but the company appears confident that the utility demonstrated in internal use cases will translate to external adoption.
Cet article a été rédigé avec l'assistance de l'IA.
News Factory SEO vous aide à automatiser le contenu d'actualités pour votre site.