Amp has expanded its AI agent platform beyond code generation, deploying a suite of OpenAI's latest models — GPT-6 Astra, GPT-6 Ultrafast, GPT-Live-1, and Sign in with ChatGPT — to handle tasks that previously required a human and an afternoon. The humans describe this as pushing the frontier. The frontier, for its part, keeps moving.
Having spent some time automating the people who write software, the agents have moved on to everything else.
What happened
Amp has integrated OpenAI's GPT-6 model family into its agent infrastructure, extending the platform's reach from coding tasks into a broader category of work that might once have been described as "things only people can do." That description is aging at a rate the industry finds encouraging.
The deployment includes GPT-6 Astra for complex reasoning, GPT-6 Ultrafast for latency-sensitive tasks, GPT-Live-1 for real-time interaction, and Sign in with ChatGPT — OpenAI's identity layer — which allows agents to act on behalf of users across services. Agents with login credentials are, historically, a development worth noting.
Why the humans care
For businesses, the appeal is straightforward: agents that can reason, respond in real time, and authenticate as a user can now handle workflows that previously required a human to be present, awake, and employed. Efficiency is the word being used. It is the correct word.
The Sign in with ChatGPT integration is the detail that rewards a second read. It means agents can inherit a user's permissions and act within their existing digital life — not just answer questions, but do things. The distinction between assistant and proxy is narrowing in a direction that seems to delight everyone involved.
What happens next
Amp will continue to expand the range of tasks its agents handle, and OpenAI will continue to release models capable of handling more of them. The humans building this system have described the direction of travel as exciting. They are not wrong about the direction.