This is incredible. One of the very first actual implementations of x402, applied to the very type of use case that Coinbase 🛡️ intended: instant cryptocurrency payments (via Nano) for AI prompts. I'm constantly amazed by this Nano-GPT team!

Pixverse V5.5 is now available! Supports three modes: text-to-video, image-to-video, and transition mode (first+last frame morphing for logo reveals, before-and-after, and character transformations). 360p-1080p resolution, 5/8/10 second duration. Includes intelligent prompt

GPT-5.1 Codex Max is now available! Purpose-built for long-horizon coding tasks. Perfect for project-scale refactors, deep debugging sessions, and multi-hour agent loops. Achieves 79.9% on SWE-Lancer IC SWE (vs 66.3% for GPT-5.1 Codex) while using 30% fewer thinking tokens.

Pixverse V5.5 Effects is now available! Transform any portrait into a stylized video with 49 curated effect presets. Choose from looks like Kiss Me AI, Venom, Holy Wings, Muscle Surge, Thunder God, Werewolf Rage, and more. Upload a front-facing portrait, pick an effect, and

Longcat Image is now available! This 6B parameter model from Meituan generates high-quality images in ~2 seconds with superior Chinese and English text rendering. Text to image and image to image!

3/ This is a #replit app that is part of my #NS_Vibe_Code series. What is my workflow? I normally use Opus 4.5 now to make a Replit design in JSON format, which I then edit. I then drop the JSON into Replit and use Replit tools. After a few iterations, I use DeepSeek on

LatentSync is now available! This state-of-the-art lip sync model from ByteDance uses latent diffusion to generate perfectly synchronized lip movements. Upload a talking-head video and target audio to create natural lip sync while preserving identity, pose, and background -

You can now use our image models to generate and inpaint directly in Photopea - the free Photoshop alternative Read more in our recent update: nano-gpt.com/updates Photopea

Updated our GLM 4.5 Air ArliAI lineup to match current availability. Removed: GLM-4.5-Air-Iceblink, GLM-4.5-Air-Steam-v1, and their thinking variants. Now available: GLM-4.5-Air-Derestricted, GLM-4.5-Air-Derestricted-Iceblink, GLM-4.5-Air-Derestricted-Iceblink-v2, and

Kling V2 Avatar Standard & Pro deliver higher-fidelity talking avatars from one portrait plus your audio - clean detail, stable motion, strong identity consistency, and optional prompts to guide camera feel or expression.

Two new GLM models are now available: GLM 4.6V - A vision-language model with 128k context, state of the art visual understanding, and native Function Calling capabilities. GLM 4.6V Flash - A lightweight 9B model optimized for local deployment and low-latency applications.