Gemini Omni 1.1 Flash extends AI video to 40-second scenes
New keyframe control and a $0.03/s draft tier make it cheaper to prototype video — plus Cursor's repo-less agents and ChatGPT's multi-account Gmail.
Copy markdown
40-second scenes, built ten seconds at a time
Gemini Omni 1.1 Flash extends clips in 10-second steps to 40s total, now reading up to 10 seconds of prior footage (not just the last frame) so motion and style carry across the cut. You can also feed in up to 3s of your own video as a style reference.
Keyframe control — and a 4K that's upscaled
Set explicit start and end frames to steer camera moves between keyframes via the image_to_video task. Outputs run 360p through 4K, but 1080p and 4K are upscaled from a lower native render — worth knowing before you promise a client 'native 4K.'
The new price sheet: draft at $0.03 a second
A 360p draft mode runs about 60% faster at a third the cost — $0.03/s versus $0.10 (720p), $0.15 (1080p), and $0.30 (4K). Rough it out at 360p, then re-render the keeper. Model ID gemini-omni-1.1-flash, live in Google AI Studio; the old preview endpoint sunsets Sept 30.
Cursor: cloud agents that start with no repo
Cursor's Cloud Agents can now begin with no connected GitHub or SCM — it auto-creates an 'Origin' repo, port-forwards the live environment straight to your browser, and can publish to a live URL via Vercel. A genuine zero-setup weekend build.
ChatGPT: two Gmail accounts in one chat
ChatGPT and Codex now connect multiple Gmail, Calendar, and Contacts accounts and use them in the same conversation — personal and work side by side. Rolling out globally on Plus, Pro, Business, and Enterprise across web, desktop, iOS, and Android.