Latest news
Turn an Architectural Render into a Video: A Step-by-Step Guide to Gemini OmniRendering Architecture with Unreal Engine Path Tracer: A Step-by-Step GuideArchitectural Image Generation with Nano Banana: Prompting and Editing GuideMechaPaint 2.0 Adds PBR Painting and Photo-to-Geometry ToolsVeras 5.2 Launches: Editable 3D Models from References with GPT-6 AstraBlender Android Support: Version 5.3 Demonstrated on a Wacom TabletDesign Museum Gent Renovated with a White Waste-Based Brick ExtensionCinema 4D 2026.4 adds a built-in MCP server and tool permissionsGemini 4 Argon announced: 1 million-token output and phased accessChaos Vantage 3.4 released: OpenPBR, Chaos Scans and DLSS 4.5 arrive
AI

AI Models

Image, video and 3D generation models and the tools built on them.

Transformation of reference sketches into an editable contemporary interior modelAI image
AI Models

Veras 5.2 Launches: Editable 3D Models from References with GPT-6 Astra

Chaos announced Veras 5.2 on 29 September. Its new Agent mode can create an editable 3D scene from up to 12 image references and transfer the result to Revit, Rhino or SketchUp.

1 October 20262 min read
A contemporary design studio filled with architectural drawings and material samplesAI image
AI Models

Gemini 4 Argon announced: 1 million-token output and phased access

Google announced Gemini 4 Argon as a new model capable of generating up to 1 million tokens of output. Access will roll out in phases; introductory API pricing is $2 per million input tokens and $10 per million output tokens.

1 October 20262 min read
3D reconstruction of a historic stone courtyard building made of illuminated Gaussian splat particlesAI image
AI Models

AMD to Acquire World Labs: New Owner for Atlas World Model and Marble

World Labs signed an agreement to be acquired by AMD on 28 September. On 1 September, the company introduced Atlas, a model that can generate Gaussian splats from a few photos and produce one-minute, camera-controlled videos.

29 September 20263 min read
A moving-camera view through a modern villa interior, from the entrance to a terrace overlooking the seaAI image
AI Models

Kling 4.0 announced: 30-second 4K video, 10 keyframes and 15 references

Kuaishou’s video unit Kling AI announced Kling 4.0 on 27 September. The model generates clips up to 30 seconds long in 4K, and accepts up to 10 keyframes and 15 references. The Flash version entered a limited beta on 28 September.

29 September 20263 min read
A designer chair model, half shown as a quad wireframe and half textured, in a studio settingAI image
AI Models

Tripo P2.0 launches: AI-generated 3D models with native quad topology

Tripo AI released P2.0 on 21 September, a model that generates quad topology directly. It supports meshes with up to 25,000 quad faces, regional regeneration and one-click UV unwrapping; competitor Meshy focused on 4K detail with version 7.1.

29 September 20263 min read
High-resolution image of a residential complex at dusk on a large screen in a dark presentation roomAI image
AI Models

FLUX 3 Video adds 4K output: 2K and 4K support in the BFL API

On 10 September, Black Forest Labs added support for 2560×1440 and 3840×2176 output to the FLUX 3 Video API. 4K image-to-video generation costs $0.80 per second; high resolution is an upscaled version of the same generation.

28 September 20263 min read
A desk scene where a pencil sketch of a living room becomes a photorealistic interior renderAI image
AI Models

ChatGPT Images 2.5 Launches: Sketch, Focused Editing and Two New API Models

OpenAI made ChatGPT Images 2.5 available to all users on 8 September. Highlights include Sketch for generating images from drawings, regional editing and up to 50% lower latency; the API offers two models, Flare and Sunburst.

27 September 20264 min read