Augment Code is an AI coding platform built around a codebase context engine. It ships as extensions for Visual Studio Code and JetBrains IDEs, plus Auggie CLI for terminal workflows and Cosmos for cloud-agent sessions. The main product surfaces are Agent, Chat & Agents, Auggie CLI, MCP/native tools, usage analytics, Cosmos, and AI pull-request review. Code completions are now a buyer caveat because Augment’s feature-availability docs label them as Enterprise and say non-Enterprise completion support on the former Indie, Standard, Max, and Legacy plans was deprecated on March 31, 2026.
The positioning is different from Cursor or Windsurf: Augment is not trying to replace your editor. It tries to understand a large repository deeply enough to make code changes, answer questions, and review pull requests without forcing a new IDE.
The model menu matters. As of June 22, 2026, Augment documents Claude Fable 5, Claude Opus 4.7, Opus 4.6 and 4.5, Sonnet 4.6 and 4.5, Haiku 4.5, Google Gemini 3.1 Pro, OpenAI GPT-5.5, 5.4, 5.2, and 5.1, and Moonshot Kimi K2.6. Augment also ships Prism smart-routing presets for Claude+Gemini and GPT+Kimi. Treat the Fable 5 listing as a model-menu claim, not a guaranteed production route. Anthropic separately says Fable 5 and Mythos 5 access is suspended for all customers as of June 12, so teams should confirm the actual model route in their Augment workspace before basing rollout decisions on Fable behavior.
System Verdict
Pick Augment Code if your problem is codebase context. The strongest fit is a professional team with a large existing repo, multiple IDE preferences, and a need for coding agents plus PR review under one paid contract.
Skip it for simple autocomplete or beginner app building. GitHub Copilot is cheaper for completion-heavy work. Cursor and Claude Code are better-known choices for developers who want a full agentic coding environment. Replit Agent, Lovable, and Base44 are better for non-developers building apps from prompts.
Who pays which tier: Business at $100/month is the current public default for teams up to 50 seats. It includes $100/month of usage across LLM tokens service fee, and Cosmos compute, plus pay-as-you-go after included usage. Enterprise is the custom route for more than 50 seats, custom usage, SSO/OIDC/SCIM, CMEK, ISO 42001, SIEM, data residency, granular access controls, audit trails, and dedicated support.
Key Facts
| Core product | Codebase-aware AI coding platform |
| Interfaces | VS Code extension, JetBrains extensions, Auggie CLI |
| Main features | Agent, Chat & Agents, Auggie CLI, MCP and native tools, usage analytics, Cosmos, Code Review |
| Agent scope | Can create, edit, or delete files and use terminal and tools with reviewable diffs and checkpoints |
| PR review | Available on all paid plans; Enterprise adds advanced review controls |
| Selectable models | Augment docs list Claude Fable 5, Opus 4.7/4.6/4.5, Sonnet 4.6/4.5, Haiku 4.5, Gemini 3.1 Pro, GPT-5.5/5.4/5.2/5.1, Kimi K2.6, Prism routing |
| Cosmos | Included in Business and Enterprise; compute billed at $0.19/hour in 5-minute increments when used |
| Billing unit | Dollar-denominated usage balance across LLM tokens, a 40% LLM service fee, and Cosmos compute |
| Paid data use | Paid plans exclude AI training on customer data under commercial terms |
| Pricing | Business $100/month flat for up to 50 seats with $100 included usage; Enterprise custom |
What It Actually Is
Augment is an assistant layer for developers who already live in an IDE. Agent handles multi-step coding tasks. Chat answers repo questions and helps plan changes. Auggie CLI brings the same context engine into terminal sessions, while Cosmos moves longer sessions into Augment-managed cloud-agent environments. Treat completions as an Enterprise entitlement check rather than a safe Business-plan assumption.
Agent can plan and implement features, upgrade dependencies, document changes, queue tests in the terminal, open Linear tickets, and start pull requests. The important practical detail is reviewability: Augment shows code diffs, tool calls, terminal commands, and checkpoints so a developer can steer or roll back a run.
The product is strongest when the repository already has meaningful structure: tests, package scripts, conventions, docs, and review habits. Augment’s context engine can surface that structure to the model; it cannot manufacture engineering discipline from an untested codebase.
Decision Matrix
| Need | Augment fit | Why |
|---|---|---|
| Existing enterprise codebase | Strong | Editor extensions, context engine, paid no-training terms |
| Greenfield app from a prompt | Weak | Replit Agent, Lovable, or Base44 are more direct |
| Multi-editor team rollout | Strong | VS Code, JetBrains, and CLI surfaces |
| Heavy autonomous CLI loop | Medium | Auggie CLI exists, but Claude Code is cleaner for terminal-first work |
| Low-cost autocomplete | Weak | GitHub Copilot is cheaper and simpler |
| PR review plus coding | Strong | Code Review can use the same paid usage balance |
When To Pick Augment Code
- You have a large monorepo or long-lived product codebase. Augment’s pitch is repository understanding, not greenfield app scaffolding.
- Your team uses multiple editors. VS Code, JetBrains, and CLI coverage make it easier to standardize the assistant without forcing one editor.
- You want coding and review in one tool. The same Business usage balance can cover Agent work, Cosmos, CLI sessions, and Code Review.
- You care about commercial data terms. Paid plans state that customer data is not used for AI training.
- You want MCP and native tools. Augment can connect to external tooling for richer context and task execution.
When To Pick Something Else
- Best GUI-first AI IDE: Cursor. Stronger all-in-one agent window and editor-native orchestration.
- Best terminal agent: Claude Code. Cleaner for autonomous CLI loops.
- Cheapest mainstream coding assistant: GitHub Copilot. Better value if completions are the main need.
- Open-source, bring-your-own-key agent: Cline or Continue.
- Non-developer app builder: Lovable, Base44, or Replit Agent.
Pricing
Pricing via Augment Code pricing:
| Plan | Price | Included usage | Fit |
|---|---|---|---|
| Business | $100/month flat | $100/month across LLM tokens, service fee, and Cosmos compute; up to 50 seats | Small and mid-size teams that want Augment without per-seat math |
| Enterprise | Custom | Custom usage, custom top-up, unlimited users, custom compute, multi-region compute | Larger teams that need SSO/OIDC/SCIM, CMEK, ISO 42001, SIEM, data residency, access controls, audit trails, and dedicated support |
at the provider’s public API at $0.19/hour billed in 5-minute increments. Once the included $100 is consumed, billing continues on pay-as-you-go at the same rates with no minimum top-up. Pricing verified 2026-06-22 via augmentcode.com/pricing and Augment token-based pricing.
Against The Alternatives
| Augment Code | Cursor | GitHub Copilot | |
|---|---|---|---|
| Primary shape | IDE extensions + CLI | Full VS Code fork | Extensions across popular IDEs |
| Best moat | Codebase context engine | Agent window and editor integration | Distribution through GitHub/Microsoft |
| Team fit | Mixed-editor teams | VS Code-first teams | Broadest mainstream adoption |
| Pricing floor | $100/mo Business | Free, then $20/mo Pro | $10/mo individual Copilot |
| Review workflow | Built-in Code Review | Bugbot add-on | GitHub-native review features |
| Best viewed as | AI layer for production codebases | AI-native editor | Default commodity coding assistant |
Failure Modes
- Usage accounting is real work. Teams need to watch usage by mode, model, user, and component because Business spend is a shared dollar balance across LLM tokens, the 40% LLM service fee, and Cosmos compute. Fable, Opus, GPT-5.5, and Cosmos-heavy runs can consume the included $100 quickly.
- Not a full editor replacement. That is a feature for some teams, but developers who want a purpose-built AI IDE may prefer Cursor or Windsurf.
- Agent quality depends on repo hygiene. Large codebases with weak tests, sparse conventions, or fragile build steps still need careful human review.
- The old solo/team tiers are stale. The current public pricing surface no longer shows Indie, Standard, or Max as the primary buyer paths. Solo developers who expected a $20/month plan should verify account-level availability before budgeting.
- Enterprise value depends on controls. SSO, CMEK, ISO 42001, data residency, SIEM integration, and audit trails matter only if your organization will use them.
- Model menus change. Augment’s current docs list Fable 5, Opus 4.7, GPT-5.5, Kimi K2.6, and Prism, but availability and defaults can shift. Because Anthropic separately says Fable 5 access is suspended, verify the exact model route in the workspace before using Fable-specific results as rollout evidence.
- Code completions need an Enterprise check. Augment’s feature-availability docs label code completions as Enterprise and say support was deprecated on former non-Enterprise plans. If completions are the reason to buy, confirm the entitlement in the account before rolling out Business.
Methodology
This page was produced by the aipedia.wiki editorial pipeline. Scoring follows the four-dimension rubric at /about/scoring/ (Utility, Value, Moat, Longevity; unweighted average). Last verified 2026-06-22 against primary Augment sources: the pricing page, token-based pricing, available models, feature availability, Agent docs, and introduction docs.
FAQ
Is Augment Code free? No persistent free tier was used for this review. The current public pricing page starts at Business for $100/month flat with up to 50 seats and $100 included usage; Enterprise is custom.
Does Augment work in JetBrains? Yes. Augment documents JetBrains IDE support. Code completions specifically need an entitlement check because the current feature-availability docs label completions as Enterprise.
What is Auggie CLI? Auggie CLI brings Augment’s agent, context engine, and tools into terminal workflows.
Does Augment train on paid customer data? Augment’s pricing FAQ says paid plans exclude AI training on customer data under its Commercial Terms of Service.
Can I use Claude Fable 5 or Opus 4.7 inside Augment? Augment’s available-models page lists Claude Fable 5 and Opus 4.7 alongside Sonnet 4.6, Haiku 4.5, Gemini 3.1 Pro, GPT-5.5, and Kimi K2.6, plus Prism routing presets. However, Anthropic separately says Fable 5 and Mythos 5 access is suspended, so verify the live route in your Augment workspace before treating Fable as available.
What is Cosmos? Cosmos is Augment’s cloud-agent platform for launchable, monitorable, resumable sessions, experts, environments, integrations, webhooks, files, and cloud compute. It is included in Business and Enterprise, with compute billed at $0.19/hour in 5-minute increments when used.
Sources
- Augment Code pricing: Business and Enterprise plan packaging, included usage, seats, top-ups, and no-training statement
- Augment token-based pricing: usage components, 40% service fee, Cosmos compute, model rates, and usage dashboards
- Augment feature availability: Cosmos, clients, and code-completion entitlement caveats
- Augment documentation: enterprise coding platform, Cosmos, Auggie CLI, and SDLC automation
- Augment Agent docs: Agent capabilities, checkpoints, tool use, review flow
- Augment available models: current selectable model list and picker behavior
- Anthropic Fable/Mythos access statement: Fable 5 and Mythos 5 suspension caveat
Related
- Category: AI Coding
- Alternatives: Cursor · Claude Code · GitHub Copilot · Windsurf · Cline · Continue
