A fine-tuned variant of Qwen3.6 35B with reduced content filtering, this model takes a more permissive approach to outputs than its base counterpart. It handles both text and image inputs, making it multimodal in practice. The 'aggressive' tuning suggests guardrails have been deliberately loosened, which means it may produce content that standard deployments would decline — a trade-off users should weigh carefully.