Time to warm up the GPUs I guess, since I can’t be fucked to do it myself lol. Thank gods for qwen flash next and DeepSeek!
TBH the only reason I’m not working mainly with flash next anyway is that I can only run inference for one agent at a time on my setup.