catalog · pricing comparison

— models. Always cheaper than direct.

Every model below is priced against the home provider's public API rate for the same workload (1K-input + 4K-output ceiling for chat, per image, per 5-second 720p clip for video). All settled in USDC over x402.

grep

fetching …

how we compare

Chat is benchmarked at 1,000 input + 4,000 output tokens, which is a comfortable ceiling for most production calls. Image is per default resolution. Video is per 5-second 720p clip.

pricing sources

Anthropic, OpenAI, Google AI, xAI, Kling, ByteDance public price pages as of May 2026. Newer Opus 4.7 priced same as 4.6 per Anthropic's policy.

why cheaper

Aggregated through an OpenAI-compatible routing layer with multi-account load balancing. We pass most of that savings through; the rest covers facilitator + on-chain settlement.