DEDUPED Hot AI repos, turned into use cases Get the Monday 5

How to set reasoning effort for thinking models in Chat with Omni Copilot

Control how hard reasoning models think by choosing an effort tier in chat or in settings.

Short answer: Pick a reasoning-capable model in the Copilot Chat model picker; then if VS Code shows an effort control next to the model, choose the tier there; the choice is forwarded automatically; if no control is shown, open settings and set omnicopilot.defaultReasoningEffort to none, low, medium, high or xhigh. All steps.

Who it is forA developer who uses reasoning models such as GPT-5.x, Claude with thinking or DeepSeek R1 and wants to trade speed for depth.

The problemReasoning models respond at a fixed depth unless you set an effort level, and you want a default when the editor shows no effort control.

The outcomeYour chosen effort tier is sent to OmniRoute as reasoning_effort for models that support extended thinking.

What you need first

  • VS Code 1.104 or later, or a VS Code fork such as Cursor, Windsurf or VSCodium
  • Node.js with npm, to install the OmniRoute CLI
  • A running OmniRoute server, local or remote, with at least one provider configured or the built-in free ones

Steps

  1. Pick a reasoning-capable model in the Copilot Chat model picker.
  2. If VS Code shows an effort control next to the model, choose the tier there; the choice is forwarded automatically.
  3. If no control is shown, open settings and set omnicopilot.defaultReasoningEffort to none, low, medium, high or xhigh.
  4. Send a test prompt and confirm the response reflects the chosen level.

Prompt for your agent

Using the diegosouzapw/OmniCopilot README, set omnicopilot.defaultReasoningEffort to <effort-level> in the VS Code settings for <workspace-path>. Report which models the setting applies to. Verify that an explicit choice in the chat UI overrides the setting.

Caveat. The default applies only to models the catalog marks as reasoning-capable, so it is ignored for other models.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for diegosouzapw/OmniCopilot and "Set reasoning effort for thinking models in Chat"

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Share: X LinkedIn Hacker News Email Copy link

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer