Not budget models
The Eco pool includes models that match flagship frontier models on coding benchmarks and score near-perfect on competition math. They are not weaker; they are efficient.
Loading...
A curated pool of AI models built on sparse architectures that activate 3–6% of their parameters per response. Serious intelligence, a fraction of the compute, and a fraction of the cost and energy.
When a dense frontier model answers you, hundreds of billions of parameters activate, whether you asked for a business plan or the weather. Sparse mixture-of-experts models route your request through only the expert sub-networks relevant to the task.
The knowledge is encoded across all of the parameters: the routing is what's efficient. Less compute per response means less energy per answer and dramatically lower cost, without meaningfully sacrificing quality on everyday work.
The Eco pool includes models that match flagship frontier models on coding benchmarks and score near-perfect on competition math. They are not weaker; they are efficient.
Eco keeps the same smart routing as Auto Mode. ARMES analyzes each request and picks the best model for the task, directly from the efficient pool instead of the frontier pool.
Because each response consumes a fraction of the compute, your usage budget stretches roughly 10–30x further. If frontier routing gives you ~100 quality conversations, Eco can give you 1,000+.
Efficient AI isn't a corner of ARMES: it is woven through the platform. Choose the entry point that fits how you work.
The Eco plan: $9/mo
A full paid AI subscription built entirely on the efficient pool. ARMES GPT ECO includes Auto, Max, and Eco modes—where Auto and Eco share the same high-efficiency routing logic—or lets you pick any model in the Eco + Free pool manually. Includes Eco-tier agents, ARMES Research, unlimited notes, and meaningfully higher limits than Free.
Cancel anytime with one click in the billing portal. Compare plans
Eco Mode: on Pro & Ultra
Already on Pro or Ultra? Every agent has Eco Mode in its mode selector, right beside Auto and Max. Use Auto or Max for the work that demands the frontier; switch to Eco for drafts, brainstorms, code reviews, and daily conversations, stretching your included budget dramatically further. One app, three strategies, zero waste.
Eco Mode is included: no extra cost, no setup.
Frontier routing optimizes for absolute best intelligence. Eco optimizes for the best intelligence per dollar and per watt. Knowing when to use each is the whole trick.
You need PhD-level reasoning, complex legal or strategic analysis, formal mathematics, or multi-hundred-page document synthesis. Auto Mode dynamically pairs agile workhorses with apex escalations to right-size compute, while Max Mode bypasses heuristics entirely to lock into the top reasoning models of your subscription tier (scaling up to uncapped frontier engines on Pro & Ultra). When precision supersedes token efficiency, the frontier remains the right tool.
Quick questions. Drafting emails. Brainstorms. Code reviews. First drafts. Research summaries. Daily conversations. For the vast majority of everyday AI work, you won't notice a meaningful quality difference, while your budget and footprint stretch much further.

Not every question needs the most powerful model in the world to answer it. ARMES gives you the choice, and rewards you with far more usage when you take it.