Skip to content
Discussion options

You must be logged in to vote

Yes, but only between two levels — DeepSeek's API on v4 reasoner exposes high and max, with no low / medium rungs like OpenAI's reasoning_effort.

All three presets default to max (auto / flash / pro). To run high:

  1. Edit ~/.reasonix/config.json, add "reasoningEffort": "high"
  2. Or use /preset fast — legacy alias for flash + high effort, kept for backwards compat, not surfaced in the picker

Current effective value shows up in /status.

The default-to-max is deliberate: effort mainly affects the length of the reasoning chain (token count), not the per-token rate. For typical coding turns the high → max cost delta is small, but the accuracy impact (especially on pro-tier hard turns) is real.

Replies: 2 comments

Comment options

You must be logged in to vote
0 replies
Answer selected by dacec354
Comment options

You must be logged in to vote
0 replies
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Category
Q&A
Labels
None yet
2 participants