New Models in Rovo? (Sonnet/Opus/Fable 5, Luna/Tera/Sol)

[Jeremy Scheffel]

I’m writing to inquire about Atlassian’s planned rollout of new models being available within Rovo Dev. I ask because, as this community is likely well-aware, there are newer and more capable models from OpenAI/Anthropic that (as far as I’m aware) are not available yet on Rovo Dev. Is there a timeline for any of the recent models being available within Rovo – such as GPT 5.6, Sonnet/Opus/Fable 5?

Also – what can we expect going forward regarding model availability within Rovo? Is there a standard Atlassian clearing procedure/timeline?

[Viswanathan Ramachandran]

Hey @Jeremy Scheffel 

Great question. The gap between a model dropping and appearing in Rovo Dev usually comes down to three enterprise essentials: Zero Data Retention (ZDR) guarantees, context graph integration, and credit multiplier calibration.

Here’s the TL;DR on what to expect:

• Model Lineup: Atlassian takes a multi-model approach (OpenAI, Anthropic, Google). Rovo Dev uses dynamic routing in Auto mode to match task complexity to the right model, but you can always check or switch manually using the /models command in the CLI.

• The Clearing Pipeline: Before a new model goes live, Atlassian must ensure strict code privacy (your IP is never used for training) and test context-awareness across Jira and Bitbucket.

• Where to Track: Keep an eye on the Atlassian Cloud Roadmap (filter for app) and also subscribe to the Rovo Community Group for instant release drops. For ex: like this one.You can also use /feedback directly in your Rovo Dev CLI to vote for the models you want next!

Bottom line: The focus isn't just launching every new model instantly, but delivering the best ones safely with full context.