Changelog v4.12.1 - August 08, 2025

Major update: HTML code editing in frontend and GPT-5 model with advanced reasoning, multimodal support, and online search, now available in AI Hub and agents.

✨ Major Features & Improvements

  • Enabled HTML code editing: Users can now edit generated HTML code directly in the frontend, making quick adjustments without the need to regenerate outputs or spend additional AI credits.

  • Added GPT-5 model support: The new GPT-5 model is now available in both the AI Hub and for all agents. Key enhancements include:

    • Multi-step reasoning and improved chain-of-thought capabilities for complex tasks

    • Multimodal input handling (text, images, voice), with groundwork for future video support

    • Extended context windows for much longer inputs and outputs

    • Superior coding and debugging skills, especially for large codebases

    • More reliable, accurate, and nuanced responses, with reduced hallucinations

    • Advanced agentic task execution and tool use for real-world workflows

    • Ability for agents to search and incorporate the latest online information

Last updated