JEV-35B is a model based on Qwen3.5-35B-A3B that activates only 3 billion of its 35 billion parameters per token to function as a fast decision model. It outperforms the previous JEV-27B-VL version on the Decision Index benchmark while maintaining competitive visual reasoning performance.

Name: JEV-35B is a model based on Qwen3.5-35B-A3B that activates only 3 billion of its 35 billion parameters per token to function as a fast decision model. It outperforms the previous JEV-27B-VL version on the Decision Index benchmark while maintaining competitive visual reasoning performance.
Uploaded: 2026-10-17T20:00:38.732Z
Channel: ModelScope2022
Description: JEV-35B is a model based on Qwen3.5-35B-A3B that activates only 3 billion of its 35 billion parameters per token to function as a fast decision model. It outperforms the previous JEV-27B-VL version on the Decision Index benchmark while maintaining competitive visual reasoning performance.

ModelScope2022agent-decisions5.6K views · 95 likes

Discovered 2026-10-17, refreshed 2026-10-20. Descriptions and stats are pulled from the project's own X post and refreshed automatically — they aren't independently verified by JevTracks beyond the initial eligibility check.