Skip to main content

sonar_reasoning_pro

What it is​

A module-level ChatModel instance configured for the Perplexity provider using LangChain’s ChatPerplexity, pre-set to the sonar-reasoning-pro model.

Public API​

  • Constants

    • MODEL_ID: str — Set to "sonar-reasoning-pro".
    • PROVIDER: str — Set to "perplexity".

  • Objects

    • model: ChatModel — A ready-to-use chat model wrapper:
      • Underlying model: langchain_perplexity.ChatPerplexity
      • Configuration:
        • model="sonar-reasoning-pro"
        • temperature=0
        • timeout=120
        • api_key loaded from ABIModule.get_instance().configuration.perplexity_api_key (wrapped in pydantic.SecretStr)

Configuration/Dependencies​

  • Dependencies

    • langchain_perplexity.ChatPerplexity
    • naas_abi_marketplace.ai.perplexity.ABIModule
    • naas_abi_core.models.Model.ChatModel
    • pydantic.SecretStr

  • Required configuration

    • ABIModule.get_instance().configuration.perplexity_api_key must be set (Perplexity API key).

Usage​

from naas_abi_marketplace.ai.perplexity.models.sonar_reasoning_pro import model

# Use the configured ChatModel in your application.
# (Exact invocation depends on ChatModel's interface in naas_abi_core.)
print(model.model_id, model.provider)

Caveats​

  • Importing this module initializes the underlying ChatPerplexity client immediately and reads the API key from ABIModule configuration.
  • Temperature is fixed at 0 and timeout at 120 seconds in this definition.