Gemini Omni 1.2 Flash: 15-second video at $0.07 a second
Google's video model goes GA at 720p, reportedly ~30% under 1.1. Plus OpenAI's gpt-6-cyber, a leaner Framer Agent harness, and outbound GPT-Live calls.

Copy markdown
Native 15-second clips at 720p
Gemini Omni 1.2 Flash is generally available on the Gemini API, Google AI Studio and Gemini for business — up to 15-second 720p clips at $0.07 per second, with multi-turn conversational edits and steadier subject consistency across a shot. Reportedly ~30% cheaper than 1.1 Flash, cheap enough to spin ad and product-video variants without rationing regenerations.
OpenAI ships gpt-6-cyber
A new OpenAI model, gpt-6-cyber, is purpose-built to investigate vulnerabilities, reproduce them in an authorized test environment and verify fixes; OpenAI points you to a “Daybreak” setup guide before the first request. For security teams it's a dedicated agent model rather than a general chat model pressed into service.
Framer Agent gets Harness 2
Framer rebuilt the Agent harness so one program reads the document, decides and applies changes as the model writes its next steps — Framer claims 2x fewer tool steps and 39% fewer cached input tokens per message, and exposes the same flow to external agents over MCP. A companion Convert update lets the Agent build and analyze funnels and A/B tests.
GPT-Live can now place outbound calls
Outbound SIP calling is live for every org: connect gpt-live-1 to a phone number through the Live API with your own SIP provider. That flips voice agents from inbound-only to dialing out — confirmations, reminders and follow-up calls placed by the model itself.