OpenRouter says Jev 1.13 beats Sonnet 5 on an internal test set

OpenRouter says Jev 1.13 had the highest accuracy and lowest cost among compared Decisions models, and against Claude Code's use of Sonnet 5, on an internal test set the company generated for the task.

The company's post on X does not publish the test set, the accuracy figures or the cost figures, so the comparison cannot be checked from the report alone.