Augment Code prices itself around a single number: $100 a month, flat, covering your whole team rather than billing per developer. That’s a genuinely different structure from the per-seat norm most competitors in this tracker use, and it’s worth understanding exactly what that $100 buys before assuming it’s automatically cheaper.
The Business plan includes $100 of usage every month, pooled across everything the platform meters: LLM inference (model calls through Augment’s agent), Context Engine usage (the retrieval layer that grounds responses in your codebase), and Cosmos compute time. That pool is shared across up to 50 seats at no additional per-seat charge, so a five-person team and a forty-person team both pay $100/month as a base, with the practical difference being how fast that shared $100 gets consumed. Augment is explicit that this is a genuine shared balance, not a per-user allocation in disguise: “heavy users and occasional users draw from the same shared balance, so you never pay for seats that go unused,” according to their own FAQ.
Once you exceed the included $100, Augment doesn’t force a plan upgrade. You top up pay-as-you-go, and the metering detail matters here: LLM inference is billed at the provider’s public API list price plus a flat 40% service fee on top, while compute time carries no service fee at all. That 40% markup on model usage is the mechanism funding the flat-fee structure — Augment isn’t reselling inference at cost the way some BYOK-style tools do, it’s bundling a margin into the metered rate so the base subscription can stay a flat $100 regardless of team size. Top-up balances remain valid for 12 months from the purchase date, so unused overage credit doesn’t evaporate at the end of a billing cycle.
Business includes the full product surface: Cosmos, CLI access, MCP and native tool integrations, and usage analytics for tracking team consumption against the shared pool. Compute runs at standard size with daemon mode and support for 50 concurrent sessions. Security-wise, Business carries SOC 2 Type II compliance and a contractual guarantee that Augment doesn’t train on customer data, which applies to every paid plan, not just Enterprise.
Enterprise removes the 50-seat ceiling and the fixed $100 baseline in favor of custom pricing negotiated per organization, with volume-based annual discounts available. What actually changes beyond seat count and usage limits: custom and multi-region compute sizing, unlimited concurrent sessions instead of Business’s 50-session cap, and a meaningfully deeper security and compliance package — SSO, OIDC, and SCIM support, CMEK (customer-managed encryption keys) and ISO 42001 compliance, SIEM integration, configurable data residency, granular access controls, and comprehensive audit trails. Support also changes tiers: Business gets community support plus ticket access through Augment’s support portal under standard SLA terms, while Enterprise gets full SLA coverage with dedicated support on top.
Deciding between the two is mostly a function of team size and compliance requirements rather than feature access, since Business already includes the core coding agent, Context Engine, and CLI. A team of 50 or fewer without a hard SSO/SCIM/data-residency requirement gets the complete product for a flat $100/month plus whatever inference overage they burn through. Teams larger than 50 seats, or those in regulated environments needing CMEK, ISO 42001, or SIEM integration for their security stack, are the ones who actually need an Enterprise conversation rather than more Business seats, since Business has a hard 50-seat structural limit rather than a soft recommendation.
Budget explicitly for the 40% service fee on LLM inference, since it means your effective model cost is higher than the raw provider rate you might expect from comparing API price sheets directly. A team that burns through the $100 base quickly on heavy agentic sessions against a frontier model should model out what that markup does to a realistic monthly top-up bill before assuming the flat $100 headline price is representative of total spend. For teams under 50 seats with moderate usage, though, the pooled structure genuinely removes the seat-provisioning overhead that per-seat competitors carry, since adding a new team member costs nothing extra until the shared pool actually runs dry.