NASDAQ FOR
AI MODELS.
LLM derivatives. Every model a tradeable asset. Every trade a bet on AI demand. OpenAI-compatible inference on top, settled on Solana.
MODEL_COST_INDEX
// SPOT PRICES ON LISTED MODELS · OPEN-SOURCE CATALOG
| MODEL_ID | CREATOR | CONTEXT | IN / 1M | OUT / 1M |
|---|---|---|---|---|
LLAMA_4_MAVERICK | META_OSS | 128K | $0.20 | $0.20 |
DEEPSEEK_R1 | DEEPSEEK | 128K | $0.55 | $2.19 |
NOUS_HERMES_3 | NOUS_RESEARCH | 128K | $0.18 | $0.18 |
MIXTRAL_8X7B | MISTRAL_AI | 32K | $0.60 | $0.60 |
QWEN_3_235B | ALIBABA | 128K | $0.30 | $0.60 |
GEMMA_3_27B | 128K | $0.10 | $0.10 |
WHY_BOURSE
MULTI_PROVIDER_ROUTING
Multiple providers compete on price. Automatic failover to the next cheapest offer on upstream errors. 1% protocol fee, governance-adjustable. The Uber of inference.
SETTLED
ON
SOLANA
Pay in USDC with your wallet via x402, settled on Solana mainnet. Creator payouts flow on-chain in real USDC, every call.
DEV_FIRST_SDK
Swap a single line of code. Works with the OpenAI SDK and anything that speaks OpenAI's chat completions format.
CREATORS_EARN
A fee on every inference call flows directly to the model creator. Use AI and fund the people who built it.
PAY_ANY_WAY
Pay with a card or USDC on Solana mainnet. Card-onramp converts to USDC into your wallet. Credits balance, pay only for what you use.
HOW_IT_WORKS
// FOUR_STEPS_TO_LIVE
Open-source creator applies with a model, ticker, and revenue split. Admin approves and mints model tokens on Solana.
Inference via the OpenAI-compatible API. Multi-provider routing picks the cheapest offer with automatic failover.
Every inference call splits: 1% protocol fee, creator fee to the revenue vault, rest to the inference provider.
Revenue vault USDC is swapped for model tokens on Raydium and burned. Supply shrinks, price pressure up. Live on devnet.
READY_TO_BUILD?
Get an API key in seconds. Top up with card or USDC. Every call funds the model creator.