← All posts
September 10, 2026 · Jim

Five new models are now available in Zide

  • Models & Model Choice
  • Product Updates
The Zide Assist model picker listing the newly added GPT-6 Astra, Claude Fable 5.1, Gemini 3.8 Flash, Qwen 3.8 Flash, and DeepSeek V4.1 Flash alongside their parameters.

Zide Assist just got a lot more interesting. We have added five models from OpenAI, Anthropic, Google, Alibaba, and DeepSeek.

GPT-6 Astra is OpenAI’s new flagship for difficult coding and research. Claude Fable 5.1 is Anthropic’s latest model for long-running work that calls for a careful read of the system before the coding starts.

Gemini 3.8 Flash and Qwen 3.8 Flash add two fast, multimodal options with million-token context windows. Both are well suited to coding tasks that include screenshots, diagrams, or a large amount of project context. Qwen is also one of the least expensive models in the Zide catalog.

DeepSeek V4.1 Flash rounds out the release with image input, strong tool use, a million-token context window, and up to 384,000 output tokens. It replaces the previous V4 Flash models in our catalog.

All five connect directly to their providers.

Give them a real task

Open a repository and type /model in Zide Assist to switch between them. The conversation keeps the project context you selected, and every proposed change remains visible alongside your files and Git state.

You can also let one model write the code and ask another to review the commit with Crosscheck. Trying Astra against Fable, or Gemini against Qwen, takes a few moments and often reveals a useful difference in how they approach the same problem.

Download Zide, open a project, and see which one earns a place in your workflow.