This Free AI Just Caught The Billion Dollar Giants
Qwen 3.8 Flash Next, a free open-weights MoE model, rivals paid closed AI systems.
“it seems that we can switch out paid closed systems to open weights AI that we can download and run ourselves forever”
Alibaba's Qwen 3.8 Flash Next introduces a mixture-of-experts architecture with three technical innovations: block-based sparse attention (QSA) for cheaper long-context handling, gated residual branching to reduce layer interference, and improved token chunking. The model is free and open-weights, running at 38 tokens/second on NVIDIA DGX Spark hardware, positioning it as a viable no-subscription alternative to proprietary frontier models.