In partnership with
Hello from The AI Night,
Today in AI:
Google Launches Gemini 3.5 Flash Agentic Model
xAI Launches Grok Integration for OpenClaw
Google Launches Gemini Omni for AI Video Creation
Here's the deal: Google released Gemini 3.5 Flash, a new model built for agentic tasks and coding. It's the first in the 3.5 family, with 3.5 Pro expected next month.
The Breakdown:
3.5 Flash scores 76.2% on Terminal-Bench 2.1, 1656 Elo on GDPval-AA, and 83.6% on MCP Atlas, outperforming Gemini 3.1 Pro on agentic and coding benchmarks.
Output speed is 4x faster than other frontier models at less than half the cost.
It powers Gemini Spark, a personal AI agent rolling out to trusted testers now and AI Ultra subscribers next week.
Available today in the Gemini app, AI Mode in Search, Google Antigravity, Gemini API in AI Studio, and Android Studio.
3.5 Pro is already in internal use, with a public launch planned for next month.
The bigger picture: Google's mid-tier model just outscored its own flagship on agentic benchmarks. That collapses the pricing floor for coding agents overnight. But Gemini Spark is the bigger story. Google is shipping a personal AI agent to Search users before competitors figure out consumer distribution. OpenAI and Anthropic have better agents. Google has billions of people already typing into a search bar.
Here's the deal: xAI now lets SuperGrok and X Premium subscribers run Grok models inside OpenClaw, an open-source personal assistant that operates locally on any hardware. The integration turns Grok into a persistent agent accessible across messaging platforms.
The Breakdown:
OpenClaw runs on Mac, Linux, Windows, VPS, or even a Raspberry Pi with persistent memory across sessions.
The agent connects to WhatsApp, Telegram, Slack, Discord, Signal, and iMessage for cross-platform access.
Every SuperGrok and X Premium subscription tier unlocks Grok inside OpenClaw at no extra cost.
Setup takes a single curl command on macOS/Linux with guided onboarding and device-code auth for headless environments.
xAI says more open-source agent integrations are coming soon.
The bigger picture: Second open-source agent integration in weeks, after Hermes. The pattern is clear now. OpenAI and Anthropic spend billions building agent platforms. xAI lets others build the platform and just supplies the brain. Running on a Raspberry Pi proves this scales down to hardware no competitor is targeting. xAI is quietly assembling distribution from the edges while everyone else fights over the center.
Here's the deal: Google released Gemini Omni, a model that generates video from any mix of text, image, audio, and video inputs. The first version, Omni Flash, is live in the Gemini app, Google Flow, and YouTube Shorts.
The Breakdown:
Users edit videos through natural language, with each instruction building on the last while maintaining character and scene consistency.
The model draws on Gemini's knowledge to simulate gravity, kinetic energy, and fluid dynamics.
A digital avatar feature generates videos that look and sound like the user.
All outputs include SynthID watermarks, verifiable through the Gemini app and Google Search.
Available to AI Plus, Pro, and Ultra subscribers. Free on YouTube Shorts and YouTube Create App.
API access for developers arrives in the coming weeks.
The bigger picture: Runway charges $96/month. Sora needs a ChatGPT subscription. Google made AI video free on YouTube Shorts, where creators already publish. By the time competitors adjust pricing, millions of creators will already associate AI video with YouTube. Google is not entering this market. It is using distribution to make the competition irrelevant before they can respond.
Want to get the most out of ChatGPT?
ChatGPT is a superpower if you know how to use it correctly.
Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.
Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.
What else you need to know:
Google DeepMind expanded Project Genie to all AI Ultra subscribers globally, adding Street View grounding that lets its world model generate interactive environments based on real U.S. locations.
Google introduced Gemini Spark, a proactive personal AI agent designed to manage tasks and navigate users' digital lives, launching to trusted testers this week and US AI Ultra subscribers next week.
Cursor launched a Jira integration that lets teams assign work items to a cloud agent directly from tickets, with the agent scoping tasks from ticket context and delivering pull requests.
Google Labs introduced Gemini for Science, bundling three experimental tools for hypothesis generation, literature review through NotebookLM, and algorithm discovery powered by AlphaEvolve and its Empirical Research Agent.
Andrej Karpathy announced he has joined Anthropic to work on frontier LLM research, calling the next few years especially formative, while planning to eventually resume his education work.
That’s it for today’s edition of The AI Night.
Our goal is to cut through the noise, surface what actually changed, and explain why it matters.
2 ways to support us:
Forward this to your AI-curious friend → https://www.theainight.com
Sponsor The AI Night and reach 500+ AI builders daily → passionfroot.me/theainight