News

GPT-6 Sol and Luna are now in GoHighLevel workflows: what to do

On September 23, 2026 HighLevel added GPT-6 Sol and GPT-6 Luna to its AI Agent and GPT workflow actions, replacing GPT-5.6 at lower cost. Here is what that means for your automation.

Key takeaways

  • HighLevel added GPT-6 Sol and GPT-6 Luna to the AI Agent and GPT actions on September 23, 2026.
  • HighLevel says they replace GPT-5.6 with more accurate output and 50% lower pricing.
  • Use Luna for high-volume classification and summaries, Sol for multi-step reasoning.
  • Test on a copy of each workflow before switching the live one.

What changed in GoHighLevel on September 23, 2026?

On September 23, 2026, HighLevel made OpenAI's GPT-6 Sol and GPT-6 Luna models available in the AI Agent action and the GPT action inside GoHighLevel workflows. According to HighLevel's changelog, they replace the GPT-5.6 models, with more accurate outputs and pricing 50 percent lower.

Both models appear in the Model dropdown of the two workflow actions. Nothing changes in existing workflows until you pick the new model, so this is an opt-in upgrade. The primary source is HighLevel's changelog entry, GPT-6 Sol and GPT-6 Luna Now Available in AI Agent and GPT Actions, and OpenAI's own announcement of the models.

What is the difference between GPT-6 Sol and GPT-6 Luna?

GPT-6 Sol is the model for complex work: multi-step agent tasks, workflows that span several apps, and decisions that need careful reasoning. GPT-6 Luna is the fast, low-cost model for high-volume tasks such as classifying messages, writing summaries, tagging contacts and drafting quick replies.

Which model for which workflow task

Sol or LunaTree: Luna for tagging messages, summaries and quick replies; Sol for AI Agent actions that decide next steps and for qualification.GoHighLevel AI actionsPick per taskTag messagesLunaMulti-step agentSolSummariesLunaQualificationSol
Luna for volume, Sol for reasoning.
Workflow taskBetter fitWhy
Tag an inbound message by intentLunaSimple classification at high volume
Summarize a call or conversation into a noteLunaShort, well-defined output
Draft a first reply to a new leadLuna, or Sol for high-value leadsDepends on how much reasoning the reply needs
AI Agent action that looks up data and decides next stepsSolMulti-step reasoning
Qualifying a lead against several rulesSolFewer reasoning mistakes matter more than cost

What does this mean for your GoHighLevel automation?

For most GoHighLevel accounts, this means the AI steps in your workflows can cost less and make fewer mistakes, but only after you switch each action to the new model and test it, because a different model can phrase replies differently or classify edge cases another way.

Switching models safely

Model switchList workflows using AI actions, duplicate one and switch the copy, compare outputs, switch the live workflow.List workflows with AI actionsGPT and AI AgentDuplicate and switch the copyLuna or SolCompare outputsTags and classificationsSwitch the live workflowOnce outputs match
Test on a copy; a new model can classify edge cases differently.
  1. List every workflow that uses the GPT or AI Agent action.
  2. Duplicate one workflow, switch the copy to GPT-6 Luna or Sol, and run test contacts through both.
  3. Compare outputs, especially tags and classifications that later branches depend on.
  4. Switch the live workflow once the outputs match or improve.

If you do not know which of your workflows use AI steps, that is a sign the account needs an audit. See fix a messy GoHighLevel account.

Should you switch every AI step right away?

No. Switch AI steps that run often or that make decisions first, where lower cost or better accuracy pays off most, and leave low-volume, working steps until you have tested the new models on them.

A workflow that classifies thousands of messages a month gains most from Luna's lower cost. A workflow where an AI Agent decides whether to book, escalate or disqualify a lead gains most from Sol's accuracy. For how AI steps fit into workflows generally, see GoHighLevel automation and AI workflow automation. More updates are on the news page, or ask us on a free call via the contact page.

Related: GoHighLevel CRM, AI sales agent, and a real AI-qualified lead flow in the Ultra Cryo case study.

Want to see what this would look like in your business?

Book a free 30-minute strategy call. We look at what you run today and tell you what to build first.

Book a Free Strategy Call

Get in touch

Tell us what you want to automate or build.

Send a few lines about your business and what is not working. We reply within one business day with next steps, or book a call if you would rather talk now.

  • Free 30-minute strategy call, no obligation
  • Written plan with scope and price before any build
  • Six months of maintenance included on every build
Alpit Patel
Alpit PatelFounder, Autoesta · HighLevel Certified Admin
Book a Free Strategy Call