Cline 3.37: GPT-5.1, Nous Research Models, and Avalon Speech-to-Text

GPT-5.1 ships with Cline-optimized prompting. Plus Nous Research's Hermes 4 models, Avalon speech-to-text with Linux support, and critical bug fixes.

Cline 3.37: GPT-5.1, Nous Research Models, and Avalon Speech-to-Text

Cline 3.37 ships GPT-5.1 with tailored prompting, adds Nous Research as a provider, and introduces Avalon speech-to-text with Linux support.

GPT-5.1 ships with optimized prompting

GPT-5.1 is now available with model-specific prompting, enhanced tool usage, and deep-planning optimizations. Enable Native Tool Calling for best performance. Read our deep dive on GPT-5.1's unique approach to coding.

0:00
/0:35

Nous Research provider

Nous Research joins as a provider with Hermes 4 models (70B and 405B). These open-weight frontier models come with custom system prompts optimized for Cline workflows and prioritize neutral alignment.

Select Nous Research from your provider dropdown and choose between Hermes-4-70B for faster responses or Hermes-4-405B for maximum capability. Both models understand Cline's tool usage patterns and coding workflows through tailored system prompts.

Avalon speech-to-text with Linux support

Speech-to-text now uses Aqua Voice's Avalon model for improved transcription accuracy. Linux users can now use voice input; previously this was macOS and Windows only. Enable it in settings under Speech-to-Text.

Model-family deep-planning

The /deep-planning command now uses model-family specific prompting. GPT-5.1, Claude, and Gemini each receive optimized prompts that leverage their unique strengths for complex planning tasks.

Enterprise configuration

Organizations can configure multiple inference providers, set MCP marketplace allowlists, and enforce global rules across teams. This iterative rollout continues building toward full enterprise deployment capabilities.

Fixes

  • Duplicate tool results: Fixed existence checking to prevent duplicate executions
  • XML escaping: Resolved entity escaping issues in model content processor
  • Commit messages: Restored commit message generation in command palette
  • Temperature parameters: Fixed type conversion for OpenAI Compatible providers
  • HTTP proxy: Expanded proxy support throughout the codebase
  • Developer workflow: Added npm run dev script for streamlined terminal development

More in this release

This release also includes improved focus chain prompting for frontier models, expanded HTTP proxy support, and additional bug fixes. See the full changelog and documentation for complete details.