Tech • AI • Robotics • Game

VIDEO
ENFR
TodayPlay for freeBloxLeaderboardShortsTop StoriesFor youTopicsVideosYT channelsArchivesSearchFavorites

Gemini 4 Argon launches as OpenAI debuts Dots and Space

AISaturday, October 3, 2026· 14 videos

Briefing

Audio player
0:00 / 0:00

Google unveils Gemini 4 Argon

Google introduced Gemini 4 Argon as a new flagship model aimed at software engineering, enterprise workflows and cyber defense. The company is emphasizing long-context reasoning, coding and autonomous security work rather than a general consumer-chatbot pitch. Early positioning places Argon in direct competition with top models from OpenAI and Anthropic. The launch also signals a broader move into the Gemini 4 generation after months of quiet testing.

Argon rollout starts under lock

Google is releasing Gemini 4 Argon in phases, starting with Fairwind and a small circle of trusted cyber defenders. The company said the model can identify, validate and patch serious vulnerabilities autonomously, making misuse risk a central launch issue. A less restricted version is being reserved for vetted defenders, while broader access is promised later for API customers, enterprises and consumers. U.S. voluntary pre-release review was cited as part of the safety process.

Argon prices undercut rivals

Google set introductory pricing at $2 per million input tokens and $10 per million output tokens for Gemini 4 Argon. Cached inputs receive a 95% discount, or roughly $0.10 per million tokens, a notable incentive for repeated large-context workloads. That pricing looks designed to attract developers running heavy coding and agentic tasks at scale. The use of the word introductory also leaves room for later increases.

OpenAI launches Dots assistant

OpenAI unveiled Dots, an always-on assistant inside ChatGPT built to act proactively across connected software. The product combines a cloud browser, computer-use capabilities and links to more than 4,000 apps, with access through channels such as Slack, phone, SMS and email. Dots is positioned less like a chatbot and more like a persistent cloud worker that can monitor inputs, execute tasks and return updates. Permission controls and approval steps are being used to gate sensitive actions.

ChatGPT Space targets teams

OpenAI also introduced ChatGPT Space, a shared workspace for documents, dashboards and AI-assisted collaboration. Teams can organize work into spaces and pages while assistants generate assets, edit content and update materials directly inside the environment. The product extends ChatGPT beyond conversation toward a more native productivity layer for organizations. Together with Dots, it suggests a push to make OpenAI the operating interface for office work.

Sign in with ChatGPT shifts billing

OpenAI rolled out Sign in with ChatGPT, allowing third-party apps to run model usage against a user’s existing ChatGPT plan. The change could materially reduce inference-cost pressure for small developers that previously had to absorb or separately bill for API usage. In practice it creates a bring your own compute model, where the subscription travels with the user across compatible software. That could alter distribution, monetization and customer acquisition across the AI app ecosystem.

Miniax M3.1 Flash surprises

Miniax launched M3.1 Flash preview as a fast, low-cost multimodal model focused on coding, research and creative work. Early testing highlighted strong results in game building, interface generation and even 3D outputs despite the model’s flash-tier positioning. It supports a 512K context window with an option to extend to 1 million tokens, and access was temporarily made free through October 7 for subscribers in Miniax Code. The release underscores how aggressively smaller vendors are competing on speed and price, not just frontier prestige.

White House reframes AI race

Donald Trump hosted leading AI executives at the White House around a new superintelligence framework for U.S. policy. Participants included figures from OpenAI, Google, Anthropic, xAI, Meta and NVIDIA, signaling an effort to align industry and state messaging. The language shift from artificial intelligence to superintelligence appeared designed to elevate strategic urgency and frame competition with China. Yet the document itself appeared light on hard rules, leaving governance unsettled as FTC scrutiny and safety arguments continue.

Videos covered

Previous briefings · AI