Augment Code

Executive Summary

What it is: Augment Code sells the Cosmos platform (cloud environments, experts, automations) and Auggie CLI to teams only: Business at $100/mo for up to 50 seats and Enterprise custom, billing per token at provider rates plus a flat 40% service fee, with GPT-5.6 Sol as default and frontier models from Anthropic, Google, Z.ai, and Moonshot. Source: https://www.augmentcode.com/pricing

What to watch out for: There is no individual tier (three months running), the 40% service fee is unique among tracked suppliers, Sonnet 5 is still absent from the model table, and the company's headline 53%-cheaper-than-Claude-Code claim is self-evaluated. Source: https://docs.augmentcode.com/models/credit-based-pricing

Bottom line: Augment's August case for harness-level savings (53% per task via the rebuilt Auggie CLI harness) plus maturing Cost Analytics makes it the strongest cost-governance story for mid-size teams, if you accept unpublished quota math and the fee. Source: https://www.augmentcode.com/blog/auggie-cli-harness-rebuild-53-percent-cheaper

Key Terms

Latest Changes

Changes since the 2026-07 report.

July watch items verified:

New August changes:

Plans

Plan Price Notes
Business $100/month, up to 50 seats Per-token model pricing + 40% service fee; Cosmos compute $0.19/hour
Enterprise Custom Custom models, admin controls, Cost Analytics

No individual/free tier; unchanged since June. Source: Augmentcode – Pricing

API Pricing

Model rates (carried over, unchanged in August): Claude Fable 5 $10/$50, Claude Opus 5 and Opus 4.8 $5/$25, GPT-5.6 Sol $5/$30, Terra $2.50/$15, Luna $1/$6, GLM 5.2 $1.40/$4.40, Kimi K3 $3/$15 per MTok, all plus the flat 40% service fee. Sonnet 5 ($2/$10 at Anthropic) is still absent from the table. Source: Augmentcode – Credit Based Pricing

Model Performance / Benchmarks

Augment published no model benchmarks in August. The Auggie CLI harness rebuild post claims tasks cost 53% less than Claude Code at equivalent pass rates on Augment's internal evals (methodology in the post). Source: Augmentcode – Auggie Cli Harness Rebuild 53 Percent Cheaper

Latest News

The 53% Harness Claim (August 14, 2026)

Augment's engineering post details a rebuild of the Auggie CLI harness (prompting, context handling, tool orchestration) yielding 53% lower cost per task than Claude Code on their internal evaluation suite. It is the most explicit harness-level cost comparison published by any tracked supplier this month. Source: Augmentcode – Auggie Cli Harness Rebuild 53 Percent Cheaper

Cosmos Consolidates Integrations (August 2026)

Weeks 32-33 brought a unified Connectors hub, ClickUp/Snowflake support, automations dashboards, and maturing Cost Analytics (CSV export, per-session costs, tag filtering). Source: Augmentcode – Cosmos Week 33 Release Notes

Community Signals

No significant HN or Reddit threads on Augment were found in August. The company's own content (FDE-as-product essay; customer story of a two-person team absorbing 6x feedback volume) is the primary public signal. Sources: Augmentcode – We Think The Fde Should Be Part Of The Product So We Built One In , Augmentcode – Feedback Triage With Cosmos

Enterprise Readiness

Feature Available? Details
SSO (SAML) Yes Enterprise. Source: Augmentcode – Pricing
SSO (OIDC) Partial Hosted MCP OAuth flows documented; OIDC SSO not itemized. Source: Augmentcode – Auggie Cli 0 35 0 Release Notes
SCIM No Not documented.
Audit logs Partial Cost Analytics, session tags, workspace activity; formal audit-log export not documented.
IP indemnity No Not mentioned.
Data residency No Not documented.
HIPAA No Not mentioned.
Air-gapped / on-prem No Cloud-first (Cosmos); self-hosted agents not offered.
SLA Yes SLA and Support Policy published. Source: Augmentcode – Sla And Support Policy
Admin controls (RBAC) Yes Space/role access controls, organization-level MCP access controls, sensitive-path approval requirements, secret write-only handling. Sources: Augmentcode – Cosmos Week 32 Release Notes , Augmentcode – Auggie Cli 0 34 0 Release Notes

Transparency Gaps

Gap Details Severity
40% service fee The flat fee on all LLM usage remains unique among tracked suppliers and is not itemized per invoice. High
Individual tier absence No plan below $100/month for the third consecutive month, pricing out solo developers. Medium
53% claim methodology The Claude Code comparison uses Augment's internal evals and cost accounting; no third-party replication exists. Medium
Model table lag Sonnet 5 missing two months after launch; no published SLA for adding new models. Medium