Most AI crypto projects treat models as black boxes running on centralized GPUs. @OpenGradient flips the model: their Model Hub lets developers deploy open-source AI on-chain and earn $OPG every time someone runs inference through their model.
Think of it as an App Store economy, but for verifiable AI. Devs compete on model quality because usage = revenue. Meanwhile, HACA architecture (GPU nodes + zkML proofs + TEEs) ensures every inference is cryptographically provable — no trust required.
With 2,000+ models already hosted and 2M+ verifiable inferences, the supply side is building early. Price action is quiet right now (-3.7% 24h, volume effectively zero), which makes sense for a project whose TGE isn't until April 2026. The real question is whether Model Hub usage economics can scale before token generation.
Side note: our own algo's full open track record — wins and losses included — is in our bio if you're into that kind of transparency.
What would it take for on-chain AI model marketplaces to pull developers away from traditional ML platforms?
$OPG
#OPG
Think of it as an App Store economy, but for verifiable AI. Devs compete on model quality because usage = revenue. Meanwhile, HACA architecture (GPU nodes + zkML proofs + TEEs) ensures every inference is cryptographically provable — no trust required.
With 2,000+ models already hosted and 2M+ verifiable inferences, the supply side is building early. Price action is quiet right now (-3.7% 24h, volume effectively zero), which makes sense for a project whose TGE isn't until April 2026. The real question is whether Model Hub usage economics can scale before token generation.
Side note: our own algo's full open track record — wins and losses included — is in our bio if you're into that kind of transparency.
What would it take for on-chain AI model marketplaces to pull developers away from traditional ML platforms?
$OPG
#OPG