config.yaml

# aistudio gemini free, Rate Limits https://ai.google.dev/pricing#1_5pro
# gemini-1.5-pro: 2 rpm, 32,000 tpm, 1,500 RPD
# gemini-1.5-flash: 15 rpm, 1,000,000 tpm, 50 RPD

model_list:
# northflank
  - model_name: gpt-4o-mini-northflank
    litellm_params:
      model: openai/gpt-4o-mini
      api_base: https://yc--northflank-duckapi--b69bn8cbbs7k.code.run/v1
      api_key: os.environ/NF_API_KEY
  - model_name: gpt-4o-mini
    litellm_params:
      model: openai/gpt-4o-mini
      api_base: https://yc--northflank-duckapi--b69bn8cbbs7k.code.run/v1
      api_key: os.environ/NF_API_KEY
  - model_name: gpt-3.5-turbo-northflank
    litellm_params:
      model: openai/gpt-3.5-turbo
      api_base: https://yc--northflank-duckapi--b69bn8cbbs7k.code.run/v1
      api_key: os.environ/NF_API_KEY
  - model_name: gpt-3.5-turbo
    litellm_params:
      model: openai/gpt-3.5-turbo
      api_base: https://yc--northflank-duckapi--b69bn8cbbs7k.code.run/v1
      api_key: os.environ/NF_API_KEY

  - model_name: gemini-1.5-pro-k1
    litellm_params:
      model: gemini/gemini-1.5-pro
      api_key: os.environ/GEMINI_API_KEY1
  - model_name: gemini-1.5-pro
    litellm_params:
      model: gemini/gemini-1.5-pro
      api_key: os.environ/GEMINI_API_KEY1
      rpm: 4  # 2 * # of keys
      tpm: 6400 # 32,000 * # of keys

  - model_name: gemini-1.5-flash-k1
    litellm_params:
      model: gemini/gemini-1.5-flash
      api_key: os.environ/GEMINI_API_KEY1
  - model_name: gemini-1.5-flash
    litellm_params:
      model: gemini/gemini-1.5-flash
      api_key: os.environ/GEMINI_API_KEY1
      rpm: 30  # 15 * # of keys
      tpm: 2000000 # 1,000,000 * # of keys

  - model_name: gemini-1.5-pro-k2
    litellm_params:
      model: gemini/gemini-1.5-pro
      api_key: os.environ/GEMINI_API_KEY2
  - model_name: gemini-1.5-pro
    litellm_params:
      model: gemini/gemini-1.5-pro
      api_key: os.environ/GEMINI_API_KEY
      rpm: 4  # 2 * # of keys
      tpm: 6400 # 32,000 * # of keys

  - model_name: gemini-1.5-flash-k2
    litellm_params:
      model: gemini/gemini-1.5-flash
      api_key: os.environ/GEMINI_API_KEY2
  - model_name: gemini-1.5-flash
    litellm_params:
      model: gemini/gemini-1.5-flash
      api_key: os.environ/GEMINI_API_KEY2
      rpm: 30  # 15 * # of keys
      tpm: 2000000 # 1,000,000 * # of keys

litellm_settings:
  # Networking settings
  request_timeout: 20 # (int) llm request timeout in seconds. Raise Timeout error if call takes longer than 10s. Sets litellm.request_timeout
  num_retries: 3
  fallbacks: [{"gemini-1.5-pro": ["gemini-1.5-flash"]}]
  allowed_fails: 3 # cooldown model if it fails > 1 call in a minute.
  cooldown_time: 30 # how long to cooldown model if fails/min > allowed_fails
  drop_params: true

general_settings:
  master_key: os.environ/MASTER_KEY  # sk-1234 # [OPTIONAL] Only use this if you require all calls to contain this key (Authorization: Bearer sk-1234)

router_settings:
  fallbacks: [{"gpt-4o-mini": ["gemini-1.5-flash"]}, {"gpt-3.5-turbo": ["gemini-1.5-flash"]}]
  model_group_alias: {"gpt-4": "gemini-1.5-pro"}
  routing_strategy: simple-shuffle