3 ms·

I revisited this and using the models directly via ollama run was actually surprisingly fast. A bug or misconfiguration with the connection to opencode seemed

I revisited this and using the models directly via ollama run was actually surprisingly fast. A bug or misconfiguration with the connection to opencode seemed to be the culprit.

advanced-stack 8mo ago
mmh.. interesting; I'll try to compare the perf