Arena Combines Preference and Rubric Rewards to Post-Train FLUX.2-dev
Arena details a post-training method for text-to-image models that combines a reward model trained on 5 million human preference votes with VLM-generated rubric rewards. The approach boosted FLUX.2-dev by 69 Elo, ranking second on the T2I leaderboard.
2026-10-02 ~ 2026-10-03 · 3 related posts
- Arena's reward recipe lifts post-trained FLUX.2-dev by 69 Elo on live T2I leaderboard — arena · 2026-10-02
- Arena details post-training T2I models with 5M preference votes and VLM-built rubric rewards — arena · 2026-10-02
1 near-duplicate retellings: infwinston