Your AI team can now choose from four more models for the work in front of it. GPT-6 Astra, GPT-6 Sol, GPT-6 Luna and Claude Opus 5.5 are available for Kai Support and for Kai Code sessions in Kai Code Cloud.
A customer question and a multi-file code change ask different things of a model. This update gives your team more choice over reasoning, speed and cost while keeping customer context and code work connected in Gleap.
What each model brings
- GPT-6 Astra is OpenAI’s flagship for complex reasoning and coding. Choose it for the most difficult questions or coding tasks.
- GPT-6 Sol balances capability and cost for demanding day-to-day work.
- GPT-6 Luna is suited to repeatable, high-volume tasks where cost and throughput matter.
- Claude Opus 5.5 is suited to long, multi-step coding work, including changes that span many files. It is now the default model in Kai Code Cloud.
These descriptions follow OpenAI’s model guidance and Anthropic’s Opus 5.5 announcement. The right choice still depends on your task and the results you need to review.
Choose a model where the work happens
For customer support, choose a model for Kai Support, the AI support agent. Kai Support answers from your connected knowledge and follows the rules your team sets for replies and escalation. Model choice lets you fit that work to your volume and the complexity of the questions you receive.
For coding work, choose a model when a teammate starts a Kai Code session in Kai Code Cloud. Kai Code works from the task and its customer context, prepares a plan, changes code and opens a pull request. Your engineers review the changes, merge the pull request and ship the release.
We have set Claude Opus 5.5 as the Kai Code Cloud default. You can choose Astra, Sol, Luna or another supported model when the work calls for a different balance.
Available now
The four models are in Gleap’s model selectors for Kai Code sessions in Kai Code Cloud. Every agent is on Team, Pro and Enterprise, and Pro and Enterprise include LLM model choice. Kai Code Cloud usage is billed by the tokens it uses and the model you choose. See current plan details before selecting a plan.
Update, October 2026: Gleap now picks the model for Kai Support, which costs $0.29 per resolution. LLM model choice applies to Kai Code, custom agents and AI workflow steps on Pro and Enterprise.
Open your Gleap workspace and choose a model for the next customer answer or coding task. Your team decides what work to start and what code to ship.
