Using Mixlayer With Kilo Code

Mixlayer is an inference platform for open models such as GLM and Qwen, with a serving stack built from scratch by core contributors to Candle. It exposes an OpenAI-compatible API and is available as a built-in provider in Kilo Code.

Website: https://mixlayer.com/

Getting an API Key

  1. Sign Up/Sign In: Go to Mixlayer and create an account or sign in.
  2. Navigate to API Keys: Open the Mixlayer console and go to the API Keys page.
  3. Create a Key: Click New Key, give it a descriptive name (e.g., "Kilo Code"), and copy it. You will not be able to view it again.

Configuration in Kilo Code

Mixlayer is available as a built-in provider in Kilo Code, so you can connect it directly — no custom provider setup needed.

  1. Open Settings (gear icon) and go to the Providers tab.
  2. Click Connect provider, search for Mixlayer, and select it.
  3. Enter your Mixlayer API key.
  4. Pick a model from the list.

Models

Mixlayer serves open models including:

  • z-ai/glm-5.3 — 256K context
  • deepseek/deepseek-v4.1-flash — 256K context
  • qwen/qwen3.8-27b — 256K context, vision-capable
  • qwen/qwen3.6-35b-a3b and qwen/qwen3.5-9b — 128K context, vision-capable
  • qwen/qwen3.5-4b-free — free, 128K context, vision-capable

Tool calling and reasoning are supported across the model line. See the Mixlayer docs for the full, current model list and supported parameters.

Tips and Notes

  • Model list: The picker comes from the models.dev catalog. If a model you expect is missing, check Mixlayer's /v1/models endpoint for what your key can use.
  • Pricing: See Mixlayer pricing for current per-model pricing.
  • Reasoning: Qwen models support a thinking mode; reasoning tokens count against the output budget, so give responses enough room when reasoning is enabled.