Skip to Content
Platform FeaturesAI Model Selection

AI models

The model selector in the chat input controls which available model handles the conversation and strategy-building work. Access is controlled by subscription tier.

Available models

Display nameAccessCurrent description
GPT-5.6 LunaBasicFast responses, strong reasoning. This is the default.
GLM 5.3 FlashBasicFast, open-source model.
Sonnet 4.6BasicBalanced speed and quality.
Opus 4.8Premium and QuantAdvanced reasoning for complex tasks.

The selector shows locked models when your tier does not include them. Selecting an available model changes the model used for the chat request; it does not change the backtesting engine or the historical data.

Choose a model

  1. Open a chat.
  2. Select the model name in the chat input.
  3. Choose an unlocked model.
  4. Continue the conversation or ask Lona to build the strategy.

The default model is GPT-5.6 Luna. If a provider request cannot be completed, the application has an internal fallback path; the selector’s names and access rules are the supported user-facing contract.

Model selection tips

Describe entry rules, exit rules, indicators, parameters, position behavior, and risk constraints explicitly. Review the identified rules and generated code before running a backtest, regardless of the selected model. Model output is not a substitute for checking the implementation against your intended strategy.

Last updated on