return {
  "yetone/avante.nvim",
  dependencies = { "nvim-lua/plenary.nvim" },
  event = "VeryLazy",
  opts = {
    provider = "openai",
    openai = {
      endpoint = "https://api.openai.com/v1",
      model = "gpt-4o", -- your desired model (or use gpt-4o, etc.)
      timeout = 30000, -- Timeout in milliseconds, increase this for reasoning models
      temperature = 0,
      max_completion_tokens = 8192, -- Increase this to include reasoning tokens (for reasoning models)
      --reasoning_effort = "medium", -- low|medium|high, only used for reasoning models
    },
  },
  build = "make",
  cmd = { "AvantePrompt", "AvanteToggle" },
  keys = {
    { "<leader>aa", "<cmd>AvantePrompt<cr>", desc = "Avante Prompt" },
    { "<leader>at", "<cmd>AvanteToggle<cr>", desc = "Toggle Avante Window" },
  },
}