Models priced at $5+ per million input tokens or $15+ per million output tokens — the exact list used for DevPass fair-use limits
Premium is a pricing classification, not a curated list: a model lands on this page when any of its providers charges at least $5 per million input tokens or $15 per million output tokens. The classification is computed directly from catalogue prices, so this page always shows exactly which models are premium right now — every model not listed here is standard.
The distinction matters for DevPass coding plans, where premium models are subject to a weekly fair-use allowance (a percentage of the plan's monthly credits) on top of the normal credit balance. When you call the LLM Gateway API directly with pay-as-you-go credits, premium models have no extra cap and no markup — you pay the same per-token provider prices shown here.
Pricing alone. A model is premium when at least one of its providers charges $5 or more per million input tokens, or $15 or more per million output tokens. There is no hand-picked list — the classification is recomputed from the live model catalogue, so it updates automatically when prices change.
No. LLM Gateway charges the same per-token provider prices for premium models as for any other model, with no markup. The premium classification only affects DevPass fair-use limits — it never changes what a request costs.
DevPass plans include a weekly fair-use allowance for premium models: 12% of monthly credits on Lite, 15% on Pro, and 18% on Max. The allowance works on a fixed 7-day window that opens with your first premium request and fully resets when it ends. Standard models are never affected — they're limited only by the plan's credit balance.
If it appears on this page, it's premium; otherwise it's standard. In the full models directory, premium models are marked with a gem icon, and the DevPass models directory offers a pricing-tier filter, since the distinction only affects DevPass fair-use limits. The classification is also documented in the model categories guide in our docs.