Warp and GPT-5.5: Open Source Revolution with AI

Warp and Innovation with GPT-5.5

Warp, a modern terminal that has quickly gained popularity among developers due to its speed and collaboration features, has integrated GPT-5.5 to orchestrate agents across local, cloud, and open-source workflows. This integration allows agents to manage complex tasks while reducing token usage by 30% compared to its predecessor, GPT-5.4. Warp has seen its terminal become a natural hub for developers working with agents, where commands, context, collaboration, and review already intersect.

When Warp made its terminal client open source this year, with OpenAI as the founding sponsor of the repository, the company also introduced Open Agentic Development. This model allows for open software building, where humans define objectives and oversee outcomes, while agents plan the work, write code, test changes, and open pull requests. Recent improvements in cutting-edge AI models have made this type of agent orchestration practical at scale.

Massive Adoption and Increased Efficiency

Today, Warp is used by nearly 1 million developers and is adopted by over 56% of Fortune 500 companies. In Warp's engineering organization, agents now co-create about 90% of the company's pull requests, providing the team with direct insight into what long-running agent workflows need to evolve: observability, coordination, memory, and human review. Zach Lloyd, CEO of Warp, stated: “We believe we can ship a better Warp, faster, by working with our community to oversee a fleet of agents. OpenAI's models help make this sustainable for the long-term coding work these systems require.”

The Next Generation of Collaborative Software Development

Open Agentic Development is Warp's bet on the future of software development. Agents will write code, while developers will specify intent, verify results, and decide what will ultimately be shipped. These choices become a reusable context for future agents, allowing the system to improve over time. Warp believes that if orchestration is good enough, agents can produce more consistent code than a poorly coordinated human group. Open source then becomes less about human contribution to implementation work directly, and more about contributing to the judgment produced and the shared vision that only humans can provide.

Persistent and parallelized agents require components such as shared memory, reproducible environments, evaluation systems, permissions, and means to coordinate work. Warp has built Oz, its cloud orchestration platform, to manage agents across local and cloud environments.

Agentic Orchestration with Oz

Oz acts as a control plane to deploy and coordinate agents across local and cloud environments. Developers can launch agents via a web interface, select predefined skills and environments, choose the model and hosting configurations, and monitor long-running workflows centrally during execution. Once launched, agents can continue to operate remotely while developers inspect live sessions, monitor execution status, review generated artifacts, and exchange workflows between cloud and local environments without losing context. Oz also supports recurring workflows, allowing agents to function like scheduled cron jobs.

As agents accumulate more state over time, maintaining focus and preserving important decisions becomes increasingly challenging. Oz employs techniques such as context compaction, persistent memory, and dedicated sub-agents for tasks like code searching and file analysis to help agents remain reliable across extended workflows. OpenAI's models play multiple roles within Oz. For the Warp agent, tasks are categorized by type and difficulty, with more complex coding and reasoning work directed towards more powerful model configurations. GPT-5.5 is part of the mix of OpenAI models that Warp uses for demanding agentic coding workflows. Warp also utilizes OpenAI models as LLM-in-judge systems in its evaluation pipelines.

Building the Infrastructure for Agentic Development

For Warp, Open Agentic Development and the Oz orchestration platform are ultimately part of the same long-term bet: that software development will evolve from individual interactions with coding assistants to systems coordinating a large number of persistent agents over time. So far, this bet seems to be paying off. Warp's annual recurring revenue grew 35 times last year, with enterprise revenue up over 500% since the fourth quarter of 2025. The company claims that a significant portion of this growth comes from organizations seeking more flexible ways to scale agent workflows.

The underlying workflows around agentic development are still early and highly experimental. By making its terminal client open source and building in public with Open Agentic Development workflows, Warp hopes that developers will help shape the evolution of orchestration, oversight, and verification systems as agents become more autonomous over time. Zach Lloyd states: “No one knows exactly what the future of agentic development will look like. We believe the community should be able to participate in defining it.”