Warp has come a long way since it initially set out to modernize the terminal. In the screenshot above, an agent is working through a plan alongside a developer's own shell commands — running its own commands, reasoning, proposing a diff — all in the same scroll stream. Five years ago, none of that ...

With today’s open-sourcing of Warp, our goal is to create a new way of building, where humans and agents collaborate in the open to ship better software, more quickly. We are creating a flywheel where a user’s idea becomes a real improvement to Warp—prototyped with an agent, refined and approved wit...

Today we are announcing a fundamental change in how we build Warp: the Warp client is now open-source, and the community can participate in building it using an agent-first workflow managed by Oz, our cloud agent orchestration platform. OpenAI is the founding sponsor of the new, open‑source Warp rep...

Claude Opus 4.7 is now available in Warp on paid plans and is the default model for auto (genius). Opus 4.7 brings stronger performance to the coding and agent workflows many Warp users rely on most, and we're excited to make it available to users. In our testing, it especially stood out on tougher ...

...

The consensus amongst engineering leaders I talk to is that they want to deploy cloud coding agents to automate development this year. Their goal is for agents to be first-line contributors across the entire SDLC: from planning and prototyping to coding and testing to deploying and monitoring. The b...

Imagine you're on your phone and a teammate messages you: the onboarding flow is broken after the latest deploy, the signup form isn't validating inputs, the progress bar is stuck, and the confirmation screen doesn't render. You fire off a message to your agent from Slack: "Fix the onboarding flow. ...

This past week, the Warp team met in Tahoe for our semi-annual offsite. It's a time for the team to meet in person as a fully-remote company, to bond over ideas, and above all: to build together. This year, we held a single-day hackathon to build something with our Oz cloud agent platform. We had fu...

Today we are excited to launch Oz: a cloud-based platform for running, managing and orchestrating coding agents at scale. Oz is the easiest way to go all-in on cloud coding agents, without spending time and effort building agent scaffolding. Oz makes it easy to: Launch parallel cloud coding agents t...

We’ve optimized Warp's agent to use MCP more efficiently. MCP usage in Warp now uses 26% fewer tokens for tasks that rely on MCP context, and 10% fewer tokens when MCP context is supplied but unused. And these improvements were made for every AI model we support, without sacrificing quality. Here's ...

2025 was the year Warp stopped being “just a terminal” and became an Agentic Development Environment (ADE): a platform where the default unit of work shifted from ‘open file, type code’ or ‘open terminal, type command’ to write a ‘prompt, steer agents, and ship with confidence’. The numbers tell par...

We’re excited to announce that Warp now supports the latest OpenAI Codex models, unlocking the coding performance improvements that these models deliver over the core GPT models. Why now? Support for Codex has been a frequently requested feature since the Codex models were first released (see Warp i...

Back when we announced Warp 2.0 (hard to believe it’s only been 4 months), one of the core ideas was that agents let you multithread yourself – a single developer might have four or five different agents working on various features at once. Once you reach that level of parallelism though, it becomes...

Today we’re excited to roll out full support for OpenAI’s GPT-5.2 across Warp. After extensive benchmarking, tuning, and real-world stress testing, GPT-5.2 is now available to all Warp users, and it consistently delivers the best end-to-end coding performance we have seen yet. Alongside the release,...

Warp now supports one-click MCP install for a simplified installation flow for MCP. You can install MCP servers from: Shared MCPs from your team: Anyone on your team can publish an MCP server, and Warp automatically redacts secrets before sharing. Teammates (or your second laptop) can install it ins...

Anthropic’s latest model, Claude Opus 4.5, is now available for all Warp users. Claude Opus 4.5 is built for long-horizon tasks, which pairs perfectly with Warp’s new planning entrypoint: /plan. Creating a persistent plan gives Claude Opus 4.5 a structured place to document its reasoning and decisio...

Warp’s new features expand what agents can do and how you can interact with them: Full Terminal Use lets the agent use the terminal as you would: interacting with REPLs, debuggers, and full-screen apps like top. Warp is the only product on the market with Full Terminal Use capabilities. /plan allows...

Google’s latest model, Gemini 3 Pro, is now available for all Warp users. Give it a try, and let us know what you think. Gemini 3 Pro impressed the Warp team right out of the gate. It performed well on benchmarks and real tasks— more on both below. Over 15% improvement on Terminal-Bench 2.0 Warp’s l...

We’re excited to roll out support for OpenAI’s latest model, GPT-5.1, inside Warp. After weeks of testing, benchmarking, and tuning, GPT-5.1 has proven so capable that we’ve made it the default model for all new users starting today. Under the hood of GPT-5.1 GPT-5.1 builds on what we liked about GP...

Today we’re updating how Warp’s paid plans work. We are moving to a simpler structure: one single, $20 per month plan, plus usage-based credit reloads and bring-your-own-key (BYOK) for folks who want more AI. No bones about it: this plan will be more expensive for some users and less expensive for o...

Warp’s latest release includes full UI zoom, one of our most requested features ever. We sat down with founding engineer and Team Lead Aloke to talk about why it took so long, what finally clicked, and how a couple of flights (and Warp’s agentic coding features) led to one of Warp’s most satisfying ...

We want every Warp user to get the most out of our agents. As part of that, we are always trying to make our agents more efficient, but also want to provide users with clear data on credit usage so they can build intuition and use our agents most effectively. Whether you’re running quick commands, e...

Varoon is a rising Senior at Georgia Tech, where he studies AI and systems. Working on and with startups gives him purpose: he’s interned at several in the past, and is always tinkering on a new project. He enjoys running, reading, writing, and cooking. When I joined Warp as an intern in May 2025, w...

Today, we’re launching Warp Code — a suite of features for shipping agent-generated code all the way from prompt to production. With Warp Code you get: Top-rated coding agent: #1 on Terminal-bench (52%) and top three on SWE-bench Verified (75.8%, scored with GPT-5) as of Sep 2nd 2025. We built the U...

Our approach We built on the same production-grade, single-agent system we used in our previous SWE-bench Verified submission. The agent and tools are the same ones we ship in Warp, which keeps our results representative of real-world usage. We used a single‑agent architecture powered by GPT‑5 as th...

Note (Oct 9, 2025): We’ve updated some content in this post to reflect our recent product terminology change from “requests” to “credits.” To learn more about this update around AI usage transparency, see our blog post on making AI usage more transparent. Whether you’re a long-time Warp fan, or you’...

Today, Warp is excited to roll out support for OpenAI’s newest model, GPT-5, to all users. Internal testing shows GPT-5 to be especially strong at making changes that cut across multiple layers of the tech stack. Check out the video below to see GPT-5 tackle a real-world change in Warp, adding the a...

Over the past few months, we’ve seen Warp users push the boundaries of agentic development: launching over 2 million agents per day and writing more than 100 million lines of code each week using the #1 coding agent. From debugging massive monorepos to running complex multi-agent workflows, develope...

One month ago we launched Warp 2.0, the first Agentic Development Environment. It was born from a simple observation: development is shifting from coding by hand to coding by prompt, and existing products miss the mark in supporting this workflow. What’s needed is a product native to this agentic wo...

In this post, we walk through how to get started with Warp. Watch the below videos for an example of how you can use Warp to solve a real bug in Warp. This guide is tailored towards people who are new to coding in Warp. It offers a broad overview of some of our coding features— dive into the docs to...