config.yaml

providers:
  openai:
    api_base: https://api.openai.com/v1
    api_version: '2023-05-15'
    api_key: OPENAI_API_KEY
  mistral:
    api_base: https://api.mistral.ai/v1
    api_version: '2023-10-26'
    api_key: MISTRAL_AI_KEY
models:
  gpt-3.5-turbo:
    provider: openai
    fallback:
      - mistral-tiny
    cost_per_1k_tokens:
      input: 0.0015
      output: 0.002
    rate_limit: openai-limit
  gpt-4:
    provider: openai
    fallback:
      - mistral-tiny
    cost_per_1k_tokens:
      input: 0.03
      output: 0.06
    rate_limit: openai-limit
  mistral-tiny:
    provider: mistral
    fallback: []
    cost_per_1k_tokens:
      input: 0.00015
      output: 0.00075
    rate_limit: mistral-limit
  mistral-small:
    provider: mistral
    fallback: []
    cost_per_1k_tokens:
      input: 0.0006
      output: 0.0018
    rate_limit: mistral-limit
  mistral-medium:
    provider: mistral
    fallback: []
    cost_per_1k_tokens:
      input: 0.0027
      output: 0.0081
    rate_limit: mistral-limit
  mistral-large:
    provider: mistral
    fallback: []
    cost_per_1k_tokens:
      input: 0.008
      output: 0.024
    rate_limit: mistral-limit
rate_limits:
  default:
    requests_per_minute: 5
    tokens_per_minute: 100000
  openai-limit:
    requests_per_minute: 5
    tokens_per_minute: 40000
  mistral-limit:
    requests_per_minute: 5
    tokens_per_minute: 150000
users:
  user1:
    rate_limit: default
    total_cost: 0.00253995