Augment Code cuts GPT-5.6 Terra and Luna token rates up to 80%

Augment Code’s flat $100/month Business plan (up to 50 seats) is unchanged, but the published per-model token rate table at docs.augmentcode.com/models/token-based-pricing moved this week. GPT-5.6 Luna — the cheapest routed model, used for small-scale automations — fell from $1.00/$6.00 to $0.20/$1.20 per million input/output tokens, an 80% cut and the largest single-model rate move since Augment began publishing per-model pricing. GPT-5.6 Terra, the mid-tier “everyday coding” model, fell 20% from $2.50/$15.00 to $2.00/$12.00. Cache read/write rates fell proportionally on both models.

Alongside the rate cuts, Augment added Claude Sonnet 5 and Claude Opus 4.5 to the routed model roster at the same price as their existing siblings, and renamed “Prism (GPT + Kimi)” to “Prism (GPT),” dropping Kimi K2.6 from that routed family in favor of GPT-5.6 Sol and GPT-5.6 Luna. Augment’s illustrative “typical $100 month” usage-mix example also shifted from $60 LLM tokens / $24 service fee / $16 Cosmos compute to $70 LLM tokens / $28 service fee / $2 Cosmos compute — still summing to the same $100 included balance.