albedo-qwen3.6-35b-8d2c738d

A third-party Hugging Face deployment variant of Alibaba's Qwen3.6-35B-A3B, a sparse mixture-of-experts open-weight model with 35B total parameters but only ~3B active per token. Suited for agentic coding, long-context reasoning, and vision-language tasks. Source: supabase.