Oh My Pi routes across 40+ providers by intent (default, smol, slow, plan), so the practical question is which model to put in each slot. Its 32-tool surface — LSP, debuggers, subagents — rewards models with strong tool use. Claude Sonnet 4.6 is the quality default; DeepSeek V4 Pro, Qwen3.7 Max, and Kimi K2.7 Code cover the value slots.
The quality default — its tool-use strength gets the most from omp's LSP and debugger integration.
The value champion for the default slot — near-frontier coding at open-weight prices.
Frontier-class with a 1M context — a strong pick for the slow/plan slots on big refactors.
Coding-tuned and cheap — ideal for the smol slot handling quick edits and subagent tasks.
A well-priced GPT option if you prefer the OpenAI family across slots.
The model picks the moves; the agent runs the loop, the tools, and the guardrails. Once you've chosen a model, see which agent gets the most out of it.
Every model on this page is included with Standard Compute
