Google Drops Gemini 4 Lineup, Cheaper Video AI, and Smart Robot Brains That Actually Work

Google rolled out Gemini 4 Hydrogen on November 12, aimed squarely at speed and resource efficiency while packing a million-token context window. Priced at $1.50 per million input tokens and $7.50 for output, it features four reasoning levels and teams up with Antigravity to dynamically split chores with its heavier sibling. In automated benchmarks, this tag-team approach slashed quota usage by 24% while speeding through tasks nearly a fifth faster. Meanwhile, the heavyweight Gemini 4 Carbon stepped out of limited preview with an introductory API price of $2 per million input tokens before doubling up after November 2026. This flagship model coordinates parallel agent swarms inside isolated workspaces, letting different digital entities tackle separate parts of a single project simultaneously. On the creative front, Gemini Omni 1.2 Flash stretched video generation limits up to 15 seconds while trimming generation costs by 30% down to seven cents per second for 720p footage. The updated model introduces multi-step conversational editing that tweaks backgrounds and lighting without requiring a full re-render, alongside tighter lip-syncing for audio prompts. Shifting from pixels to metal, Google DeepMind unleashed Gemini Robotics 2.5 to handle fragile objects and resist adversarial jailbreak prompts with up to double the success rate of previous iterations. In a live factory test at Mercedes-Benz, an Apollo 2 humanoid built by Apptronik successfully sorted automotive components 96% of the time using a dual-model setup.

Another week, another frantic corporate scramble to convince everyone that chatty software and dexterous tin cans will definitely solve all productivity bottlenecks before the next subscription fee hike.

Source: BotHub