feat(ai): enable Mistral prompt caching
This commit is contained in:
@@ -948,7 +948,7 @@ async function loadModelsDevData(): Promise<Model<any>[]> {
|
||||
cost: {
|
||||
input: m.cost?.input || 0,
|
||||
output: m.cost?.output || 0,
|
||||
cacheRead: m.cost?.cache_read || 0,
|
||||
cacheRead: m.cost?.cache_read ?? (m.cost?.input ? roundCost(m.cost.input * 0.1) : 0),
|
||||
cacheWrite: m.cost?.cache_write || 0,
|
||||
},
|
||||
contextWindow: m.limit?.context || 4096,
|
||||
|
||||
Reference in New Issue
Block a user