Skip to content

The fastest, cheapest, most efficient open-source agent harness. Run more than 500 agents on a $10 VPS.

refactor(inference): improve workload tier model lookup

Steven Enamakel3cf72030f2a7
diff --git a/crates/openhuman-core/src/inference/turn_models.rs b/crates/openhuman-core/src/inference/turn_models.rs
index 86d2349302..d718006eb8 100644
--- a/crates/openhuman-core/src/inference/turn_models.rs
+++ b/crates/openhuman-core/src/inference/turn_models.rs
@@ -172,15 +172,18 @@ fn build_turn_models_crate(
.get("summarization")
.cloned()
.or_else(|| local.model_for(role)),
- tinyagents_harness::host::ModelPurpose::WorkloadTier(route) => local
- .role_models
- .get(route)
- .or_else(|| {
- local
- .role_models
- .get(route.strip_prefix("hint:").unwrap_or(route))
- })
- .cloned(),
+ tinyagents_harness::host::ModelPurpose::WorkloadTier(route) => {
+ let role = crate::inference::provider_model::role_for_provider_model_for(
+ extensions,
+ &provider_route,
+ route,
+ );
+ local
+ .role_models
+ .get(route)
+ .or_else(|| local.role_models.get(role))
+ .cloned()
+ }
});
#[cfg(any(test, feature = "e2e-test-support", feature = "rss-bench"))]
let injected = injected