Cline：CLI v3.0.74:模型请求最多重试3次,原为5次

模型请求最多重试3次而不是5次;连接在模型仅流式输出推理后断开的回合,现在会重试,而不是结束运行。

Cline

事件类型: 功能更新
版本: 3.0.74
官方发布: 2026-10-26
来源: github.com/cline/cline/releases/tag/cli-v3.0.74
本站整理: 2026-10-27

本次变化

共 5 条要点

打开已在另一个Cline应用中运行的任务时,现在会连接到该运行,而不是启动无法停止的第二个运行;cline hub status和cline doctor在Hub冻结时不再挂起。
官方原文 · c1 c2
Opening a task that's already running in another Cline app now attaches to that run instead of starting a second one that couldn't be stopped.
官方原文 · c1
`cline hub status` and `cline doctor` no longer hang on a Hub that has frozen
官方原文 · c2
模型请求最多重试3次,而不是5次;模型仅流式输出推理后连接断开的回合,现在会重试,而不是结束运行。
官方原文 · c3 c4
Model requests retry at most 3 times instead of 5
官方原文 · c3
A turn whose connection drops after the model has streamed only reasoning (for example, "The socket connection was closed unexpectedly") is now retried instead of ending the run
官方原文 · c4
Kimi K3现在为Extra High和Max使用max推理设置,而不是它拒绝的xhigh设置;Gemini API密钥(gemini提供方)不再发送原生Google Search。
官方原文 · c5 c6
Kimi K3 now gets `max` reasoning for Extra High and Max, instead of an `xhigh` setting it rejects
官方原文 · c5
Gemini API keys (the `gemini` provider) no longer send native Google Search
官方原文 · c6
对于上下文上限未知的模型,自动压缩不再运行;run_commands收到空命令列表时现在返回错误,而不是静默成功。
官方原文 · c7 c8
Auto-compaction no longer runs for models with an unknown context limit
官方原文 · c7
`run_commands` with an empty command list now returns an error instead of a silent success
官方原文 · c8
模型目录新增Liquid Inference和Munito AI提供方,并移除CrofAI、iFlow和Llama。
官方原文 · c9
Adds the Liquid Inference and Munito AI providers and drops CrofAI, iFlow and Llama
官方原文 · c9

变化明细

下表字段直接取自官方原文，保持原文语言，不做翻译。

编号	主体	动作	对象	数值	判断
c1	Cline app	attaches to	task already running in another Cline app instead of starting a second one that couldn't be stopped
c2	cline hub status and cline doctor	no longer hang	on a Hub that has frozen		否定
c3	Model requests	retry	at most 3 times instead of 5	最多 3 times; 精确 5 times
c4	turn whose connection drops after the model has streamed only reasoning	is retried	instead of ending the run
c5	Kimi K3	gets max reasoning	for Extra High and Max instead of an xhigh setting it rejects
c6	Gemini API keys (gemini provider)	no longer send	native Google Search		否定
c7	Auto-compaction	no longer runs	for models with an unknown context limit		否定
c8	run_commands	returns an error	with an empty command list instead of a silent success
c9	Model catalog refresh	adds and drops providers	adds Liquid Inference and Munito AI, drops CrofAI, iFlow and Llama