AI Agent Index
ByHeather MacAvelia·Independently reviewed·Published Jul 10, 2026·Updated Sep 24, 2026
Independently verified against live vendor data on Sep 24, 2026.

Multi-model coding agent and dev environment from Amp Frontier Corporation that runs agents in cloud orbs from the web, Mac, iOS or CLI. Free Hobby tier, Individual from $20/mo (USD).

How we scored it

Autonomy

4/5

Integrations

4/5

Pricing clarity

4/5

Evidence

4/5

Setup

4/5

The facts

Amp is a coding agent and development environment from Amp Frontier Corporation, which spun out of Sourcegraph as an independent company in December 2025. Each thread can run in an orb, a cloud machine that Amp sets up with the project's code, tools and plugins, so the agent keeps working after the developer closes the laptop. Developers start and follow threads on the web with no install, in the macOS and iOS apps, or in the terminal through the Amp CLI, whose execute mode also runs Amp from scripts and CI. Amp is multi-model. The Dial offers four modes, low, medium, high and ultra, each pairing a main model with an oracle, and in September 2026 they ran on models including GPT-5.6, GPT-6 and Claude Fable 5.1 alongside a fast GLM model, with the models behind each mode open to customization. Specialist subagents handle code search and other focused jobs. Amp connects to GitHub through a GitHub App that clones private repositories, pushes branches, opens pull requests and reads CI status, works in Slack through its Puck assistant, acts as an MCP client for local and remote MCP servers, and extends through plugins, skills and TypeScript and Python SDKs. Developers can also run agents on their own computers with runners instead of orbs. Pricing is in US dollars. The Hobby tier is free, includes every product feature, and lets developers use a ChatGPT or other model subscription or their own API keys with no Amp token fees or limits, paying for orbs as they go or running agents on their own runners at no charge. The Individual tier is $20 per month on Megawatt, with 45,000 minutes of orb time, or $200 per month on Gigawatt, with 480,000 minutes, and both include Amp credits that work on any model. Teams costs nothing extra: members stay on Hobby or a paid tier and share threads, portals, plugins, skills and pooled credits, and SAML or OIDC SSO is included once one member is on a paid tier. Enterprise is priced by quote and adds SCIM, spend limits, audit logs, network and MCP access controls, IP allowlisting and regional endpoints. Students and teachers can get Megawatt for $10 per month. Model usage beyond a tier's included credits is billed separately, either by the model provider when developers bring their own subscription or keys, with that provider's charges and limits applying, or through paid Amp credits charged at the providers' API prices with no markup for individuals and non-enterprise workspaces. Purchased credits expire twelve months after purchase, included usage resets each month without rolling over, and web search and other tools also draw on credits. Orbs are billed by the minute, from $0.08 an hour for the smallest size to $2.13 an hour for the largest, which is limited to Gigawatt and Enterprise, and larger orbs use a tier's allowance faster. Paid tiers are billed monthly. Amp suits developers and teams who want agents that run tasks in the cloud and across devices while choosing their own models and paying for model use through subscriptions they already hold. Developers who want an agent inside an AI-first code editor should consider Cursor, developers who want Anthropic's models bundled into one subscription should consider Claude Code, developers who want OpenAI's own agent within a ChatGPT plan should consider OpenAI Codex, and developers who want a free, open-source agent they run themselves should consider OpenCode. Current state Q3 2026: On September 13, 2026 Amp introduced the free Hobby tier and the no-extra-charge Teams tier, removed Amp token fees on bring-your-own subscriptions and keys for everyone outside Enterprise, and extended its zero and minimal data retention policy to all users and workspaces. Monthly subscriptions launched in July 2026, and the macOS and iOS apps followed in August. In September Amp added customizable modes and runners that can serve many directories and Git worktrees and be shared with a workspace. Amp is SOC 2 Type II certified, and neither Amp nor its subprocessors train models on customer data unless a user opts in, which a workspace admin must approve and Enterprise workspaces cannot enable.

Pricing

freemium · $20/mo

View pricing ↗

Segment

All sizes

Setup

easy

Verified

Sep 24, 2026

Transparency

Mostly Public

Contract

Month-to-month

Data training

Not Trained

Human in loop

Optional

Capabilities

agentic-codingterminal-agentmulti-file-editingautonomousgit-nativecode-generation

Pros & Limitations

Editorial assessment

Pros

  • ✓Each thread can run in a cloud orb set up with the project's code, tools and plugins, so agents keep working after the laptop closes, and threads can be started or followed from the web, Mac, iPhone or the CLI.
  • ✓Multi-model by design: four modes route work across OpenAI, Anthropic and fast models, and on Hobby, Individual and Teams developers can bring a ChatGPT subscription or their own keys with no Amp token fees.
  • ✓Team features come at no extra charge, including shared threads, multiplayer, shared portals, plugins and skills and pooled credits, and a GitHub App lets agents open pull requests and read CI status.

Limitations

  • –Model usage beyond a tier's included credits is paid separately, through the developer's own provider subscription or keys or through Amp credits at provider API prices, and purchased credits expire after twelve months.
  • –Paid tiers are billed monthly and orb time is metered by the minute: Megawatt's 45,000 minutes go faster on larger orbs, and the largest orb size is limited to Gigawatt and Enterprise.
  • –SCIM, audit logs, IP allowlisting and regional endpoints are on the quote-based Enterprise tier, and SAML or OIDC SSO needs at least one workspace member on a paid tier.

Technical Details

Deployment
webcloudclidesktopmobile
Model architectureMulti-model: four modes route across OpenAI GPT models (GPT-5.6, GPT-6), Anthropic Claude (Fable 5.1) and fast GLM models, with bring-your-own subscriptions and API keys
Avg setup timeSelf-serve signup on the web with a free Hobby tier and no sales contact. Amp does not publish a setup time.
Autonomous rateHigh within a thread. Amp runs multi-step coding threads with subagents in cloud orbs or on the user's own runners and keeps working after the user disconnects, while developers review the resulting changes and pull requests. Amp does not publish an autonomous completion rate.
Integrations
GitHubGitSlackMCP
Security
SOC 2 Type IIGDPRCCPA

Similar agents

Rating

4.0/ 5
★★★★★

Editorial score

How we score this →

Recognition

Listed 2026
Embed these on your site →

Built for

DevToolsEnterpriseStartups

Leave a review

Never displayed publicly.

Agent Stacks

See workflow stacks that feature Amp.

Compare

Related Content

Is this your tool?

Claim this listing free to get your embeddable badges, a review collection link, and the Claimed badge on your listing.

Claim listing →

Free · Every Two Weeks

AI Agent Price & Rating Tracker

Price changes, new agent launches, acquisitions, and rating updates across the AI agent market. Verified against live vendor data, not vendor marketing.

No spam. Unsubscribe anytime. We never share your email.