Qwen3.6-35B-A3B
Open-weight sparse MoE multimodal model with strong agentic coding and Apache 2.0 licensing
No preview available
About Qwen3.6-35B-A3B
Overview
Qwen3.6-35B-A3B is an open-weight multimodal model from the Qwen team, released with 35B total parameters and 3B active parameters in a sparse MoE design. Its public release includes an official Qwen announcement plus independently accessible Hugging Face and ModelScope model pages, which makes it suitable to catalog as a standalone model entry rather than only as a news item. The model is positioned around strong agentic coding, long context, multimodal reasoning, and flexible deployment through common inference frameworks.
Key Features
- Publicly released as open weights with Apache 2.0 licensing.
- Backed by official Qwen, Hugging Face, and ModelScope distribution pages rather than only a launch article.
- Combines multimodal reasoning with coding-oriented positioning and long-context support.
- Sparse MoE design aims for stronger quality at lower active parameter cost.
Kanban app for orchestrating coding agents across isolated worktrees
Agentic IDE for flow-state coding with Cascade and terminal-aware edits
Google's most capable and creative AI assistant.
AI voice generation platform for text-to-speech and voice cloning
AI-powered video generation and editing platform
Open-weight MoE flagship models with 1M context and frontier-class reasoning