Marcel

first look at DeepSeek V4.1 Flash-1017 on a massive medieval voxel world. tested it through deepseek harness using the api on max reasoning with a pretty challenging prompt. > over 2 million voxels in a single file > took around 25 minutes > cost about $5 in api usage the v4.1 flash upgrade is looking good so far. I’ll see how it holds up on other tests what do you guys think, cook or nah?