RefinedNeuro/ qwen3.8-27b-noslop

4,467 5 days ago

Plain-spoken Qwen3.8-27B fine-tune: 42% less slop, 73% fewer em dashes, same coding-agent pass rate. MTP speculative decoding built in.

tools thinking

6 models

qwen3.8-27b-noslop:latest
830840d053da • 17GB • 256K context window •
Text input • 5 days ago
qwen3.8-27b-noslop:q2_k_m
988d5b23bed0 • 11GB • 256K context window •
Text input • 5 days ago
qwen3.8-27b-noslop:q2_k_xl
ca4135552421 • 9.8GB • 256K context window •
Text input • 5 days ago
qwen3.8-27b-noslop:q3_k_xl
24e798e80c83 • 13GB • 256K context window •
Text input • 5 days ago
qwen3.8-27b-noslop:q3_k_m
ec18f40d45eb • 14GB • 256K context window •
Text input • 5 days ago
qwen3.8-27b-noslop:q4_k_m
830840d053da • 17GB • 256K context window •
Text input • 5 days ago