Gemini 4 Argon opening to everyone leads the Oct 14 AI briefing: in the past 24 hours Google moved its frontier model from early access to general availability, Apple unveiled the HomeView smart home display, NVIDIA readied the software ecosystem for RTX Spark laptops launching Oct 16, and LangSmith and ComfyUI shipped new tools for agent debugging and video post-production.
Gemini 4 Argon now generally available: intro pricing at $2 per million input tokens
Google announced on Oct 13, 2026, on its official blog that Gemini 4 Argon is now generally available. Developers and enterprises can call it via the API in AI Studio and the Gemini platform, Google AI Ultra subscribers can use it in Antigravity, and Ultra and Pro subscribers can use it in the Gemini app and AI Mode in Search. Introductory pricing is $2 per million input tokens and $10 per million output tokens, with cached input 95% off. Google says the model ranks first on the Vals Index, which measures performance across finance, coding, legal and tax work; that ranking is Google's own claim. Teams can now test it on their real workloads instead of waiting for an invite.
Apple HomeView: a 7-inch smart display for $299, on sale Nov 13
Apple introduced HomeView in an Oct 13, 2026 newsroom release: a 7-inch Liquid Retina display on a domed speaker base for $299. Pre-orders opened the same day in the US, UK, Canada, Australia, India and New Zealand, with availability from Nov 13; Singapore pre-orders start Oct 15. It recognizes up to six household members by face and switches to each person's content, and it runs a new Siri AI powered by Apple Intelligence for camera views and control of lights, thermostats and locks. Whether it is worth buying depends on real-world sound and Siri AI performance after Nov 13; the price and dates apply to those launch markets.
NVIDIA RTX Spark ecosystem ready: laptops arrive Oct 16
NVIDIA announced on its official blog that the RTX Spark software ecosystem is ready, with laptops launching on Oct 16, 2026, offering up to 128GB of unified memory and 1 petaflop of AI compute aimed at local agents. The Nemotron 3.5 Bolt 74B model, built for local agents, arrives Oct 15: 74B total parameters with about 9B active, quantized to run in roughly 24GB. The PAIR personal AI router leaves beta and adds llama.cpp support, and a faster DLSS 5 model lands with the launch. Buyers should watch first-wave pricing and thermals on Oct 16 rather than judging on paper specs alone.
LangSmith AI Search: query agent traces in plain language
LangChain launched AI Search in LangSmith on Oct 13, 2026. It is in public beta on LangSmith Cloud in GCP US, EU and APAC regions at no additional cost, and org admins can enable it in settings. Instead of pre-building evaluators or paging through traces, you describe the behavior you want, such as an agent claiming a task was done when it was not, a frustrated user, or a possible PII leak, and a fast decision model reads each trace trajectory and returns matches. Agent teams can try a few real questions on historical traces to judge recall before adopting it for daily debugging.
ComfyUI gets NVIDIA VFX nodes: upscaling, frame generation and SDR-to-HDR locally
The Comfy team released NVIDIA VFX nodes on Oct 13, 2026, packaging three RTX video capabilities as nodes that run locally on an RTX GPU: super resolution up to 4x and up to 8192 px, with denoise and deblur modes; frame generation from 2x to 16x between frames; and SDR-to-HDR conversion to BT.2020 with peak luminance adjustable from 400 to 2000 nits. The suggested order is super resolution, then frame generation, then HDR, and the release also improves EXR handling and color management. Video creators on ComfyUI can update and install the nodes from the Manager to try the full chain.
