Alibaba Qwen 3.6

Qwen3.6 35B A3B is a multimodal Mixture-of-Experts large language model designed for efficient long-context reasoning and scalable deployment. As the successor to Qwen3.5 35B A3B and the first open-weight Qwen3.6 variant, it contains 35B total parameters with ~3B active, activating 8 routed and 1 shared expert from 256 per token. It features a 40-layer hybrid architecture combining Gated DeltaNet and full attention layers. The model supports a 262K token context window, extendable to 1M, and integrates a 27-layer vision encoder for unified multimodal processing.

Qwen3.6 27B is a dense multimodal transformer model designed for high-performance agentic coding and long-context reasoning. It features 27B parameters with 64 layers, 5,120 hidden size, and 24 attention heads with 4 key-value heads, combining linear and full attention. The model supports a 262K token context window, extendable to ~1M tokens, and includes a 27-layer vision encoder. Built as the first open-weight Qwen3.6 dense variant, it delivers strong coding performance without MoE complexity.