BPFlow AI / News

GPT‑6 Cyber for security research, Gemini Omni 1.2 Flash and Google Vids

Issue date:

This issue covers a new model for security researchers and two ways to use Google’s video generation: a developer API and the Vids application.

This issue may include important events published earlier.

News archive →
01

OpenAI introduces GPT‑6 Cyber for security research

The model is designed to find vulnerabilities, assess their impact and help develop fixes. It is initially available through the API and Codex to a select group of approved Daybreak Red customers.

Key facts

  • OpenAI reports that GPT‑6 Cyber scored 48.5% on ExploitGym, compared with 32.1% for GPT‑5.6 Cyber.
  • OpenAI says researchers used the model to identify 81 previously unknown Firefox vulnerabilities classified by Mozilla as high severity and develop working exploits for 948 Linux kernel vulnerabilities. According to OpenAI, Mozilla has released fixes for many of the Firefox issues.
  • OpenAI plans to expand access to all Daybreak Red customers soon.

An objective look at what matters

Security teams could use the model to investigate code and evaluate potential fixes, but access remains restricted.

Business processes

  • Software security review — Teams could investigate vulnerabilities and develop candidate fixes.

Automation opportunities

  • Vulnerability research and patch development: The model could assist with testing and propose fixes within an authorized scope.
    Conditions: Daybreak Red access and authorization to test target systems are required.; High-risk actions need a sandbox and human oversight.

Impact on manual work

The model could take on some research tasks, but a specialist still needs to review its actions and findings.

Limitations and risks

  • The model can develop exploits; misuse and actions outside the authorized scope remain risks.
  • OpenAI requires an IP allowlist for Daybreak Red API access by default, with limited exceptions.

Sources

  1. OpenAI News: GPT‑6 Cyber: Advancing the frontier of aligned cybersecurity agents · Source date:
02

Gemini Omni 1.2 Flash becomes generally available

Google has made the model generally available through the Gemini API and Gemini for business. It generates videos between 3 and 15 seconds long and supports successive edits through text prompts.

Key facts

  • 720p video costs $0.07 per second, 30% less than Gemini Omni 1.1.
  • Generation can be guided by images, audio cues and motion clips of up to three seconds.
  • Google says outputs carry an imperceptible SynthID watermark.

An objective look at what matters

The API allows teams to integrate short-video creation into content workflows, while prompt-based edits could reduce the need to start over.

Business processes

  • Advertising and product video production — Teams could produce and adapt video variants for different placements.

Automation opportunities

  • Producing short-video variants: API calls could generate drafts from specified images and prompts.
    Conditions: Integration with the Gemini API and a cost review are needed.; Check visual details, text and brand requirements before publication.

Impact on manual work

Successive edits could reduce manual work on variants, but they do not replace editorial review.

Limitations and risks

  • Each clip is limited to 15 seconds.
  • Google says it is still improving fast sports action, rapid script changes and complex physics.

Sources

  1. Google DeepMind News: Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips · Source date:
  2. Google Blog: Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips · Source date:
03

Google Vids adds free 15-second video generation

Google has added Omni 1.2 Flash to Vids, its video creation and editing app. Anyone with a Google account or Google Workspace subscription can generate up to three 15-second videos per month at no cost.

Key facts

  • Users can upload images as references for videos in Vids.

An objective look at what matters

The ready-to-use app gives teams a way to try video generation without building their own API integration.

Business processes

  • Short informational and promotional videos — Employees could turn ideas and images into video drafts.

Automation opportunities

  • Draft creation in Google Vids: Generation from descriptions and images could simplify the initial production of a clip.
    Conditions: A Google account or Google Workspace subscription is required.; The free allowance is up to three videos per month.

Impact on manual work

Draft creation could require less manual work, but finished videos still need review.

Limitations and risks

  • The free monthly allowance limits production volume.
  • Content and visual details should be checked before publication.

Sources

  1. Google Blog: Generate longer AI videos with Omni 1.2 Flash in Google Vids at no cost. · Source date: