-
|
as title |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments
-
|
Yes, but only between two levels — DeepSeek's API on v4 reasoner exposes All three presets default to
Current effective value shows up in 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 |
Beta Was this translation helpful? Give feedback.
-
|
offer a command to adjust? |
Beta Was this translation helpful? Give feedback.
Yes, but only between two levels — DeepSeek's API on v4 reasoner exposes
highandmax, with nolow/mediumrungs like OpenAI'sreasoning_effort.All three presets default to
max(auto/flash/pro). To runhigh:~/.reasonix/config.json, add"reasoningEffort": "high"/preset fast— legacy alias forflash + high effort, kept for backwards compat, not surfaced in the pickerCurrent 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 → maxcost delta is small, but the accuracy impact (especially on pro-tier hard turns) is real.