Theo spent $1,000 benchmarking Jev Router
Theo says Jev Router matched GPT-6 Astra on low on DeepSWE, cost slightly more and took almost 5x longer to run, a day after OpenRouter launched it. He also says he loves Jev.
Built with it
Hassan open sourced the Tev1 0.8B weights
Hassan released the weights for Tev1 0.8B, the tiny Jev-like classifier he trained and filmed running locally on a Mac through Ollama at about 50ms end to end. He is blunt about the tradeoff: not as good as Tev1 4B or Jev, and it works decently well only for very simple classification.
Definitely not as good as Tev1 4B or Jev, but it runs entirely locally on a mac.
Writing criteria
Kieran Klaassen builds embeddings out of Jev answers you can read
Kieran Klaassen asks Jev a fixed set of questions about each document and uses the answers as the vector, then runs plain cosine similarity over it.
Instead of an embedding model spitting out 1,536 numbers that mean nothing, you ask Jev questions about each document.
The case against
Theo spent $1,000 benchmarking Jev Router, found it almost 5x slower
Theo benchmarked OpenRouter's new Jev Router overnight and reported performance on DeepSWE roughly the same as GPT-6 Astra on low, at slightly higher cost and almost 5x longer runtime.
Reminder that I love Jev! I use it for a ton of things and I'm super happy with it.