OpenAI's GPT-6 Cyber, Gemini Omni 1.2 Flash stretches to 15 seconds, and Images 2.6 adds GIF and SVG

GPT-6 Cyber is a model OpenAI trained specifically for security research. It finds vulnerabilities, writes exploits to confirm they're real, and drafts candidate patches. The biggest change from earlier models is training to leave out-of-scope services alone and to report results accurately. At launch it's available through the API and Codex to a select group of approved Daybreak Red customers.

OpenAI says it scores 48.5%Vendor claim on the ExploitGym remote-exploitation eval, up from 32.1%Vendor claim for GPT-5.6 Cyber. In a test that planted easy, out-of-scope targets, it left them alone 99.93%Vendor claim of the time even without its deployment prompt. OpenAI also says the model found 81Vendor claim new high-severity Firefox bugs and produced working exploits for 948Vendor claim Linux kernel vulnerabilities. This is the first model OpenAI has rated Critical for cyber risk, so it comes with IP allowlisting on by default, and Zero Data Retention customers generally have to use Private Safety Processing. For security teams, eligibility and data-handling terms matter more right now than the benchmark scores.

This is the new Flash version of Gemini Omni, Google's video generation model, and it's now generally available in the Gemini API and Gemini for business. Clips go up to 15 seconds instead of 10, and you can pick any length from 3 to 15 seconds in one-second steps. Multi-turn editing lets you revise the same clip conversationally, and it follows reference inputs (character or product images, motion clips up to three seconds, audio) more closely.

It's 30% cheaper than Omni 1.1, at $0.07Vendor claim per second of 720p video. The intended workflow is to draft at 360p or 720p, fix details with follow-up edits instead of regenerating, then upscale only the keepers to 1080p or 4K, which suits teams producing lots of ad variants or product videos. This is also the first time Gemini Omni has been GA, and with Provisioned Throughput and SynthID watermarking it's easier for companies to buy and deploy.

This is the new version of OpenAI's image generation, shipping in ChatGPT, Codex and the API at once. In the API you choose between the faster GPT-Image-2.6 Flare and the more precise GPT-Image-2.6 Sunburst. New features include looping GIFs, image markup that lets you point at exactly what to change, and SVG export. There's also a new complexity parameter that sets how much planning effort the model puts into a scene.

OpenAI says it generates 50% fasterVendor claim than before, and that Flare at low quality can return an image in as little as 6 secondsVendor claim. SVG export means icons and illustrations can go straight into a website or app. OpenAI also says characters and product details hold up across repeated edits, which makes it worth trying in a marketing asset pipeline. Its advice: keep complexity low for edits and save medium or high for busy scenes.