Home / Technology / Atlassian Deepens OpenAI Ties, Embraces GPT-6 Astra
Atlassian Deepens OpenAI Ties, Embraces GPT-6 Astra
6 Oct
Summary
- Atlassian integrates OpenAI's newest frontier models, including GPT-6 Astra.
- Partnership expands OpenAI's models into Atlassian's platform and Rovo AI.
- Atlassian maintains a model-agnostic approach, including Anthropic models.
Atlassian announced an expanded partnership with OpenAI, integrating cutting-edge models such as GPT-6 Astra and GPT-5.6 series into its platform, including Rovo AI. This collaboration connects Atlassian's enterprise context with OpenAI's ChatGPT and Codex, aiming to provide enhanced capabilities and robust governance for technical buyers using Jira, Confluence, and Bitbucket.
The integration leverages Atlassian's Teamwork Graph and introduces a rebuilt Atlassian MCP server. This server handles over 15 million daily calls, exposing more than 220 tools across Atlassian's suite. It offers enhanced governance features, including user permission enforcement, separation of actions by risk tier, and integration with Atlassian Guard for data security.
While GPT-6 Astra offers state-of-the-art performance, Atlassian emphasizes a multi-model strategy. Rovo AI dynamically routes tasks across various providers, including Anthropic, ensuring balance in capability, speed, and cost. This model-agnostic approach positions Atlassian's Teamwork Graph as the enduring asset, with AI models treated as interchangeable.
Data privacy is addressed through Atlassian's policy of OpenAI not training on customer inputs and Astra supporting Zero Data Retention for eligible API customers. Atlassian states only query-specific data is sent to OpenAI, scoped by user permissions, mitigating concerns about data access.
