While reading through the OpenGradient developer docs, one detail kept pulling my attention.
According to the documentation, verification defaults to Vanilla unless a developer actively selects another option.
That doesn't mean the network lacks stronger verification. OpenGradient supports both TEE and ZKML, and the docs clearly explain the trade-offs. ZKML can be extremely expensive computationally, while TEE is designed to cover most real-world production needs.
The design choice makes sense.
What I haven't been able to find, though, is how developers are actually using these options in practice.
@OpenGradient reported more than 2 million inferences processed by April 2026, but I haven't seen a public breakdown showing how many used TEE, how many used ZKML, and how many stayed on the default Vanilla setting.
To me, that's one of the most interesting unanswered questions around the network.
The docs tell us what OpenGradient can do. That usage data would help show what the network actually looks like in the real world.
The technology and trade-offs are clear.
The adoption patterns behind them are what I'm most curious about.
#OPG $OPG @OpenGradient
According to the documentation, verification defaults to Vanilla unless a developer actively selects another option.
That doesn't mean the network lacks stronger verification. OpenGradient supports both TEE and ZKML, and the docs clearly explain the trade-offs. ZKML can be extremely expensive computationally, while TEE is designed to cover most real-world production needs.
The design choice makes sense.
What I haven't been able to find, though, is how developers are actually using these options in practice.
@OpenGradient reported more than 2 million inferences processed by April 2026, but I haven't seen a public breakdown showing how many used TEE, how many used ZKML, and how many stayed on the default Vanilla setting.
To me, that's one of the most interesting unanswered questions around the network.
The docs tell us what OpenGradient can do. That usage data would help show what the network actually looks like in the real world.
The technology and trade-offs are clear.
The adoption patterns behind them are what I'm most curious about.
#OPG $OPG @OpenGradient