Theo slams OpenRouter for using Jev: a non-reasoning classifier that can't gauge task complexity

intellectronica · x · 2026-09-26

Theo publicly urged OpenRouter not to integrate typesafeai's Jev model. His argument: Jev is a "system one" model that categorizes rather than reasons. From a prompt alone it cannot assess real task complexity — it has zero context about the system the prompt lives in, the tools the agent can access, or the obstacles it must work around. Asked to "port this to Rust," it can't tell whether the codebase is a 100-line TypeScript file or a million-line application, nor whether a bug fix touches core complexity. That makes it unreliable for model routing in coding-agent workflows.

Original post →

More from coding & agent

coding & agent channel →