Amp expands AI agent functions beyond coding using OpenAI’s latest models

Expanding Agent Capabilities with OpenAI Models

Amp, a North‑American startup, builds AI agents that help developers turn ambitious concepts into functional software.

CEO Quinn Slack says each improvement in OpenAI models creates a new opportunity for the agent to assist, from idea selection to code generation and page design.

Amp selects and recommends OpenAI models because they account for the majority of tokens consumed on the platform.

“OpenAI is very transparent, and when there are safeguards applied, users know and they understand. That makes it really easy to use OpenAI models.

You can do all of your work in Amp with OpenAI models and be really happy.

Slack adds that users value the steady enhancements OpenAI delivers, noting, “Our users really appreciate the consistency with which OpenAI models have been getting better and better.”

Token Management and the Sign‑in‑with‑ChatGPT Feature

Heavy usage can turn token cost into a limiting factor; some Amp customers consume up to 10 billion tokens per week.

Slack argues that constant token monitoring may discourage experimentation with an agent’s full potential.

To address this, Amp introduced a “Sign in with ChatGPT” integration that grants users additional token allowances and access to newer models such as GPT‑6 Astra.

“Sign in with ChatGPT means that people can have that feeling where they’re not counting every single token,” he explains.

He further observes that the extra token budget lets users give the agent a “longer leash,” revealing capabilities that were previously hidden.

Slack reports that users now employ the agent not only for coding but also for planning, deciding what to build, and post‑deployment tasks.

New Conversational Tools and Speed Enhancements

Amp recently launched Puck, a conversational AI assistant embedded in the platform that can start and manage other agents on demand.

Powered by GPT‑Live‑1, Puck enables real‑time voice interaction, even from a smartwatch, allowing developers to monitor agents while away from their desks.

Slack envisions scenarios such as running or jogging while the assistant keeps agents active.

The company is also testing an “Ultrafast” mode with GPT‑6 Astra, which reduces latency and changes how users sequence multiple agent tasks.

Faster response times encourage more rapid switching between tasks, increasing overall workflow efficiency.

Internally, Amp’s 20‑person team relies on the agents for marketing, user onboarding, and search‑driven discovery, treating the AI as a dedicated specialist.

“We don’t have someone devoted to it, but we have an agent devoted to it,” Slack notes.

High‑quality writing and design remain essential; the team uses GPT‑6 Astra to produce documentation and page layouts that match Amp’s exacting visual and textual standards.

“We have a very unique design style and writing style. We’re very picky about this, but we’ve been really happy with how GPT‑6 Astra does.”

Slack’s personal experiments extend beyond the workplace; he fed house plans to GPT‑6 Astra, generated a 3D Blender model, and turned it into games with his children.

He recalls that “that was not possible just a couple months ago,” highlighting the rapid progress of the underlying models.

Collectively, these developments illustrate how Amp is leveraging OpenAI’s evolving suite—GPT‑6 Astra, GPT‑Live‑1, and Ultrafast mode—to broaden the scope of AI‑driven software creation.

By reducing token anxiety, adding conversational control, and accelerating response, Amp positions its agents as versatile collaborators throughout the product lifecycle.

Why This Matters

Providing developers with affordable, faster, and more interactive AI agents lowers the barrier to experimenting with complex software ideas, potentially accelerating innovation cycles.