Architect Financial Technologies has launched Liquid Inference, a real-time LLM inference marketplace that runs a live auction for every prompt. The platform allows AI providers to bid on individual requests while buyers pay the lowest qualifying offer, providing drop-in compatibility with standard OpenAI and Anthropic APIs.
Developed by the trading firm behind the AX perpetual futures exchange, Liquid Inference applies exchange-style price discovery to LLM computing. Buyers can establish specific constraints for their requests, including cost caps, time-to-first-token limits, minimum throughput, and zero data retention rules. Providers onboard through an app to monetize spare GPU capacity, managing payouts via Stripe with itemized records for every job.
The platform integrates directly with agentic coding tools like Claude Code, Cursor, and Codex. Architect is offering the first 500 users $20 in free inference alongside a referral program for account holders. This infrastructure matters because it introduces dynamic, two-sided market pricing to artificial intelligence compute resources.