qwen3.8-flash-next on a single rtx6000 (~1 euro per hour on a spot vm) with buun-llama is i think the cheapest reasoning / euro atm. hope deepseek makes me cha

qwen3.8-flash-next on a single rtx6000 (~1 euro per hour on a spot vm) with buun-llama is i think the cheapest reasoning / euro atm. hope deepseek makes me change my setup again