Jonathan Franklin
Blog
Same model, different harness: benchmarking qwen3.8-27b with Agent Bench An early Agent Bench comparison of pi, oh-my-pi and OpenCode using one locally served qwen3.8-27b model, five Python tasks and independent hidden tests.
Tell me what is slowing the business down, what you have tried, and what a successful outcome looks like. I'll help you identify the most sensible next step.