Franixo now uses task-specific AI model routing across its analysis and discovery workflows.

Instead of relying on a single model for every task, Franixo automatically routes different types of work to the OpenAI model best suited to the level of reasoning and validation that work requires.

GPT-5.6 Luna powers Brand Analysis, grounded Competitor Discovery, and Prompt Generation. Competitor discovery uses Luna's web-grounded research together with canonical-domain verification, so a suggested competitor is checked against real evidence before it reaches you.

GPT-5.6 Terra supports recovery and evaluation workflows where additional reasoning and validation are required, including parts of the platform that still rely on our established evaluation approach, such as Free Audit.

GPT-5.6 Sol is reserved for high-confidence adjudication in the workflows that need a final decision between competing signals.

This release also brings a set of accuracy improvements:

• Task-specific model selection across Franixo's AI workflows
• Better web-grounded competitor discovery
• Stronger canonical competitor-domain verification
• Fewer incorrect competitor matches caused by third-party profile and platform pages being treated as a competitor's own website
• More accurate buyer-intent detection when selecting prompts for monitoring

The result is a more reliable AI pipeline, with each task matched to the appropriate level of reasoning.