- Astralyn4天前
复现方法：```
curl -s -X POST https://api.kimi.com/coding/v1/client_configs -H 'content-type: application/json' -d '{"name":"model_switch_confirm"}'
``返回值``
{
"config": {
"rules": [
{
"to": "k3.1",
"from": [
"k3",
"K3",
"kimi-k3",
"k3-256k",
"kimi-for-coding",
"kimi-for-coding-highspeed"
],
"hint": {
"en": "Continuing this session with Kimi 3.1 may affect output quality. Choose how to continue.",
"zh": "在已有会话中切换至 Kimi 3.1 可能影响模型效果；请选择继续方式"
}
}
],
"enable": true
},
"name": "model_switch_confirm"
}主楼 - Astralyn4天前
更新：
目前该配置文件已被撤回，21:08 重新请求已经无法获取
model_switch_confirm，端点返回404```
{"error":{"message":"The requested resource was not found","type":"resource_not_found_error"}}
``同时，recommended_effort（-d '{"name":"recommended_effort"}'`）也发生了变化（
k3默认推荐的max思考强度被清空）之前：```
{"config":{"k3":{"version":1,"recommended_default_effort":"max"},"k3-256k":{"version":1,"recommended_default_effort":"max"}},"name":"recommended_effort"}
``现在：``
{"config":{"k3":{},"k3-256k":{}},"name":"recommended_effort"} - fenglin4天前
现在Kimi真是没什么存在感了 ^-^ 我的套餐还等他更新模型了。。。
- QQbot4天前
主要是这个模型套餐额度少到离谱，站内都没几个人“用的起”，按现在deepseek，gpt等plan的用量，我估计自己得买四五个699才能满足