Databricks Integrates GPT-5.5 for Enterprise AI
Databricks has partnered with OpenAI to integrate GPT-5.5 into its Data Intelligence Platform, and the rollout will reach enterprise customers in the coming weeks. The integration operates through the Unity AI Gateway, which governs access and usage across enterprise environments. As a result, organizations can deploy advanced AI capabilities within a controlled framework.
GPT-5.5 represents a major update, as it marks the first fully retrained base model since GPT-4.5. In addition, the model focuses on agentic enterprise tasks, complex document reasoning, and long-horizon coding workflows. It also powers OpenAI’s Codex agent, thereby extending its role in software development and automation.
Performance Gains Across Enterprise Workloads
Databricks evaluated GPT-5.5 using its OfficeQA benchmark, and the model demonstrated measurable improvements. When provided with correct documents, it scored 64.66 percent, compared to 57.14 percent for GPT-5.4. Therefore, the results reflect a notable gain in document-based reasoning tasks.
In end-to-end agent workflows, GPT-5.5 achieved 52.63 percent, while GPT-5.4 reached 36.10 percent. Consequently, the model reduced errors significantly in tasks that require retrieval, parsing, and computation. Moreover, broader benchmarks indicate strong performance across technical and knowledge-based evaluations.
The model achieved 82.7 percent on Terminal-Bench 2.0, which measures command-line workflows. It also reached 58.6 percent on SWE-Bench Pro for resolving GitHub issues. Furthermore, it scored 84.9 percent on GDPval, which evaluates performance across multiple knowledge-work domains.
Expanding Partnership and Competitive Landscape
The integration builds on a multi-year partnership between Databricks and OpenAI, which began in 2025. Under this arrangement, customers can access OpenAI models through SQL or APIs without moving data or deploying separate infrastructure. As a result, enterprises benefit from simplified integration and stronger data governance.
Databricks has also expanded the Unity AI Gateway to support agentic workloads. This includes governance over model interactions, access controls, and full auditing of AI processes. Therefore, GPT-5.5 deployments operate within a structured and secure environment.
Meanwhile, the broader AI market continues to evolve rapidly, as multiple companies release competing models. Recent launches include new systems from Anthropic and expanded enterprise offerings from Microsoft. Consequently, GPT-5.5 enters a competitive landscape where performance, scalability, and governance remain key differentiators.
