AI News Today (6 Nov 2026): Google Releases Gemini Omni 1.2 Flash With 15-Second Video Clips

Today's briefing highlights major updates across AI video generation, enterprise infrastructure, financial tools, and new foundational models. Stay informed on the latest tools and policy shifts shaping the AI landscape.

1. Google Releases Gemini Omni 1.2 Flash With 15-Second Video Clips

Google released Gemini Omni 1.2 Flash, a video generation and editing model featuring native 15-second clips and multi-turn conversational editing. The update also includes a 30% price cut, bringing costs down to $0.07 per second of 720p video.

Why it matters: Developers can leverage cheaper, faster video generation options with native editing capabilities for multimedia applications.

Read the full story at MarkTechPost

2. Baidu Releases ERNIE 5.5 Flash Base Model

Baidu has launched ERNIE 5.5 Flash Base, a 280B sparse mixture-of-experts base model featuring 14B active parameters. It was trained on 33.5T tokens and published prior to any mid- or post-training stages.

Why it matters: Open base models offer builders raw architectural weights to conduct customized fine-tuning for specialized enterprise tasks.

Read the full story at MarkTechPost

3. Claude Gains Personal Finance Analysis via Plaid Integration

Anthropic launched a new Personal Finance tool in Claude that allows users to connect bank and card accounts through Plaid. Users can now ask natural language questions regarding their spending habits and financial data.

Why it matters: Developers building consumer finance tools can study how advanced LLMs handle secure third-party financial data integrations.

Read the full story at TechCrunch

4. Perplexity Launches Always-On 'Teammates' AI Agents for Work

Perplexity is releasing workspace-focused AI agents called Teammates designed to take on multiple roles within a company. These agents pull memory, existing projects, and skills from existing Perplexity tools you already use.

Why it matters: Businesses gain another workflow automation option that integrates internal company memory directly into everyday tasks.

Read the full story at The Verge

5. Dropbox Acquires Tigris Data for $350M to Bolster AI Storage

Dropbox has acquired Tigris Data for $350 million to strengthen its underlying storage infrastructure. The transaction is set to support the upcoming DBX Agent Cloud storage platform built for developers.

Why it matters: Founders building heavy AI workloads can expect specialized, scalable storage solutions tailored for autonomous agents.

Read the full story at TechCrunch

6. Box Taps Nvidia Nemotron for Faster Content Retrieval

Box is integrating Nvidia's Nemotron to power a new agent-first content retrieval system. The system reportedly answers queries four times faster while using 20% fewer errors and 90% fewer tokens.

Why it matters: Enterprise teams can achieve higher retrieval efficiency and lower API token costs when querying large corporate document repositories.

Read the full story at TechCrunch

7. Together AI Releases Together Compiler to Speed Up Inference

Together AI released Together Compiler, an agent-driven loop designed to continuously re-optimize inference forward passes based on shifting traffic. In testing, it surpassed top InferenceX configurations by up to 2.6x TPS in under 24 hours per model.

Why it matters: AI developers can significantly accelerate inference throughput and optimize production workloads under variable traffic conditions.

Read the full story at MarkTechPost