Qwen company already rushed out a Jev competitor. No open weights yet.

A user tested a new model, "decision-model-preview," from the Qwen company, described as a competitor to Jev. Initial tests through AIHubMix, a platform the user advises against, found the model to be "garbage." A key issue identified was a strict 16-question limit, causing 400 Bad Request errors for longer Japanese sentences that generated 19-22 questions. The model currently only has a documentation page, with no official announcement or open weights.

This report offers the first user-generated performance assessment of Qwen's "decision-model-preview," unlike official announcements that typically lack such early, critical feedback.