mirror of
https://github.com/QuantumNous/new-api.git
synced 2026-03-30 00:46:42 +00:00
feat: support adaptive thinking
This commit is contained in:
@@ -62,6 +62,10 @@ var defaultCacheRatio = map[string]float64{
|
||||
"claude-opus-4-5-20251101-thinking": 0.1,
|
||||
"claude-opus-4-6": 0.1,
|
||||
"claude-opus-4-6-thinking": 0.1,
|
||||
"claude-opus-4-6-max": 0.1,
|
||||
"claude-opus-4-6-high": 0.1,
|
||||
"claude-opus-4-6-medium": 0.1,
|
||||
"claude-opus-4-6-low": 0.1,
|
||||
}
|
||||
|
||||
var defaultCreateCacheRatio = map[string]float64{
|
||||
@@ -86,6 +90,10 @@ var defaultCreateCacheRatio = map[string]float64{
|
||||
"claude-opus-4-5-20251101-thinking": 1.25,
|
||||
"claude-opus-4-6": 1.25,
|
||||
"claude-opus-4-6-thinking": 1.25,
|
||||
"claude-opus-4-6-max": 1.25,
|
||||
"claude-opus-4-6-high": 1.25,
|
||||
"claude-opus-4-6-medium": 1.25,
|
||||
"claude-opus-4-6-low": 1.25,
|
||||
}
|
||||
|
||||
//var defaultCreateCacheRatio = map[string]float64{}
|
||||
|
||||
@@ -143,6 +143,10 @@ var defaultModelRatio = map[string]float64{
|
||||
"claude-sonnet-4-5-20250929": 1.5,
|
||||
"claude-opus-4-5-20251101": 2.5,
|
||||
"claude-opus-4-6": 2.5,
|
||||
"claude-opus-4-6-max": 2.5,
|
||||
"claude-opus-4-6-high": 2.5,
|
||||
"claude-opus-4-6-medium": 2.5,
|
||||
"claude-opus-4-6-low": 2.5,
|
||||
"claude-3-opus-20240229": 7.5, // $15 / 1M tokens
|
||||
"claude-opus-4-20250514": 7.5,
|
||||
"claude-opus-4-1-20250805": 7.5,
|
||||
|
||||
@@ -6,7 +6,7 @@ import (
|
||||
"github.com/samber/lo"
|
||||
)
|
||||
|
||||
var EffortSuffixes = []string{"-high", "-medium", "-low", "-minimal"}
|
||||
var EffortSuffixes = []string{"-max", "-high", "-medium", "-low", "-minimal"}
|
||||
|
||||
// TrimEffortSuffix -> modelName level(low) exists
|
||||
func TrimEffortSuffix(modelName string) (string, string, bool) {
|
||||
|
||||
Reference in New Issue
Block a user