AI Agent News/stories/cline-4-1-27-refreshes-its-model-catalog-and-fixes-reasoning-and-tool-call-bugs.md

Frameworks

Cline 4.1.27 refreshes its model catalog and fixes reasoning and tool-call bugs

Cline version 4.1.27 adds an API key field for LM Studio servers that require authentication and marks read-only MCP tools with a badge. The release also changes default models and fixes several reasoning-effort and tool-call problems.

Source: github.com/cline/cline/releases/tag/v4.1.27

Cline 4.1.27 adds an API key field to LM Studio settings for servers that require authentication, and shows a read-only badge on MCP tools whose server declares them that way. The extension package also now includes third-party license notices.

The model catalog was refreshed with 94 providers from models.dev, and Lilac and Tempr were dropped. Default models changed for 67 other providers, mostly to Kimi K3.1 Pro, Gemini 4 Neon or DeepSeek V4.1 Flash 1017.

Fixes cover saving the None reasoning effort choice, which had left GPT-6 on Azure rejecting tool calls, and a Reflection Beam error after compaction or a model switch. read_files results also now arrive in full rather than being cut to 8k characters.

Why it matters if you build agents

Developers who rely on Cline's default model for new tasks should check which model now applies and pin one if their agent depends on it. Custom endpoints that rejected tool calls or reasoning settings may behave differently after this update.

Get stories like this every weekday

One short email with the day's agent tooling news.

Subscribe