Cline shuffles its default models again

What changed

If you never picked a model in Cline, it may be running a different one today. Cline’s release notes put it plainly: if you use a provider without pinning a model, expect a different default.

What changed

• ClinePass also adds MiniMax M3.1 Flash and DeepSeek V4.1 Flash 1017 alongside its new Kimi K3.1 Pro default.
• The CLI and Desktop notes count 56 other providers that get a new default model, drawn mostly from those same three.
• On the Cline provider, Claude Sonnet 5.5 is back at the top of Cline’s recommended list, so it’s what you get when no model is chosen.
• The provider list changes too: CommandCode, Reflection, Liquid Inference and Munito AI join in Cline’s main release (the CLI and Desktop add only Liquid Inference and Munito AI), while CrofAI, iFlow and Llama are dropped.
• Cline’s no-cost model list adds Reflection Beam, Nemotron 3.5 Lightning and Laguna S 2.1, and drops Mimo V2.6 Flash and Muse Spark 1.3 Contributor.

What to do

• Happy with your current model? Open your provider settings and choose it yourself, since only unpinned setups get the new defaults.
• Used CrofAI, iFlow or Llama? Those providers are no longer in Cline’s catalog, so pick another one.

Also fixed

• Gemini API keys (the gemini provider) no longer send native Google Search. Keys on Gemini’s no-cost tier used to fail every turn with a 429 quota error on Gemini 3+.
• Model requests now retry at most 3 times instead of 5.