litellm_config.yaml

model_list:
  # - model_name: "anthropic/*"
    # litellm_params:
      # model: "openrouter/qwen/qwen3-coder" # Qwen/Qwen3-Coder-480B-A35B-Instruct
      # max_tokens: 65536
      # repetition_penalty: 1.05
      # temperature: 0.7
      # top_k: 20
      # top_p: 0.8


# cpa in hf
# gemini-2.5-pro
# gemini-2.5-flash
# gemini-3-pro-preview
# gemini-2.5-flash-lite
# gemini-3-flash-preview
  # - model_name: gemini-2.5-pro
    # litellm_params:
      # model: openai/gemini-2.5-pro
      # api_base: https://cpa.applane.qzz.io/v1
      # api_key: os.environ/CPAHF_API_KEY
  - model_name: gpt-5.4
    litellm_params:
      model: openai/gpt-5.4
      api_base: https://cpa.applane.qzz.io/v1
      api_key: os.environ/CPAHF_API_KEY
  - model_name: gpt-5.3-codex
    litellm_params:
      model: openai/gpt-5.3-codex
      api_base: https://cpa.applane.qzz.io/v1
      api_key: os.environ/CPAHF_API_KEY
  - model_name: gpt-5.2-codex
    litellm_params:
      model: openai/gpt-5.2-codex
      api_base: https://cpa.applane.qzz.io/v1
      api_key: os.environ/CPAHF_API_KEY
  - model_name: gpt-5.2-codex
    litellm_params:
      model: openai/gpt-5.2-codex
      api_base: https://cpa.applane.qzz.io/v1
      api_key: os.environ/CPAHF_API_KEY

# curl -s http://acone:11434/v1/models | jq -r .data[].id
# glm-5.1:cloud
# deepseek-v4-flash:cloud
# kimi-k2.6:cloud
# kimi-k2.5:cloud

# deepseek-v3.2:cloud
# gemini-3-flash-preview:cloud
# gemma4:31b-cloud
# minimax-m2.7:cloud
# gpt-oss:120b-cloud

  - model_name: deepseek-v3.2:cloud
    litellm_params:
      model: openai/deepseek-v3.2:cloud
      api_base: https://newapi.applane.qzz.io/v1
      api_key: os.environ/NEWAPI_API_KEY_SVIP
  - model_name: gemini-3-flash-preview:cloud
    litellm_params:
      model: openai/gemini-3-flash-preview:cloud
      api_base: https://newapi.applane.qzz.io/v1
      api_key: os.environ/NEWAPI_API_KEY_SVIP
  - model_name: gemma4:31b-cloud
    litellm_params:
      model: openai/gemma4:31b-cloud
      api_base: https://newapi.applane.qzz.io/v1
      api_key: os.environ/NEWAPI_API_KEY_SVIP
  - model_name: minimax-m2.7:cloud
    litellm_params:
      model: openai/minimax-m2.7:cloud
      api_base: https://newapi.applane.qzz.io/v1
      api_key: os.environ/NEWAPI_API_KEY_SVIP
  - model_name: minimax-m2.5:cloud
    litellm_params:
      model: openai/minimax-m2.5:cloud
      api_base: https://newapi.applane.qzz.io/v1
      api_key: os.environ/NEWAPI_API_KEY_SVIP
  - model_name: gpt-oss:120b-cloud
    litellm_params:
      model: openai/gpt-oss:120b-cloud
      api_base: https://newapi.applane.qzz.io/v1
      api_key: os.environ/NEWAPI_API_KEY_SVIP
  - model_name: gpt-oss:20b-cloud
    litellm_params:
      model: openai/gpt-oss:20b-cloud
      api_base: https://newapi.applane.qzz.io/v1
      api_key: os.environ/NEWAPI_API_KEY_SVIP

  - model_name: gemma4-e4b-uncensored
    litellm_params:
      model: openai/gemma4-e4b-uncensored
      api_base: https://newapi.applane.qzz.io/v1
      api_key: os.environ/NEWAPI_API_KEY_SVIP

  - model_name: mimo-v2.5-pro
    litellm_params:
      model: openai/mimo-v2.5-pro
      api_base: https://newapi.applane.qzz.io/v1
      api_key: os.environ/NEWAPI_API_KEY_SVIP

general_settings:
  # master_key: "sk-1234" # Disabled for easy testing
  # master_key: ${LITELLM_MASTER_KEY}
  master_key: os.environ/LITELLM_MASTER_KEY

# litellm --config parent_config.yaml --detailed_debug/--debug
# os.environ["LITELLM_LOG"] = "INFO" os.environ["LITELLM_LOG"] = "DEBUG"
# https://docs.litellm.ai/docs/proxy/config_settings
litellm_settings:
  # set_verbose: true  # deprecated, use `os.environ['LITELLM_LOG'] = 'DEBUG'
  request_timeout: 180
  drop_params: true

router_settings:
  model_group_alias:
    gpt-4o: gpt-5.3-codex
    gpt-4.1-mini: gpt-5.3-codex
    gemma4:31b: gemma4:31b-cloud
    minimax-m2.7: minimax-m2.7:cloud
    deepseek-v3.2: deepseek-v3.2:cloud
    minimax-m2.5: minimax-m2.5:cloud
    gemini-3-flash: gemini-3-flash-preview:cloud
    gpt-oss:120b: gpt-oss:120b-cloud
    gpt-oss:20b: gpt-oss:20b-cloud
    # gpt-5-codex: claude-4.5-sonnet
    # gpt-5.1-codex: claude-4.5-sonnet
  # List[Dict[str, List[str]]]: Fallback model for all errors
  fallbacks:
    - gpt-5.4:
      - gemma4:31b-cloud
      - gemma4-e4b-uncensored
    - mimo-v2.5-pro:
      - gemma4-e4b-uncensored

# curl http://127.0.0.1:7860/router/settings -H "accept: application/json" -H "x-litellm-api-key: %LITELLM_MASTER_KEY%" | jq .

# need database for this to work
# curl -X 'POST' \
#   'http://127.0.0.1:7860/fallback' \
#   -H 'accept: application/json' \
#   -H 'x-litellm-api-key: sk-...' \
#   -H 'Content-Type: application/json' \
#   -d '{
#     "model": "gpt-5.3-codex",
#     "fallback_models": ["gpt-5.4"],
#     "fallback_type": "general"
# }'