Open Evaluation Agent: Efficient and Promptable Evaluation of Visual Generative Models

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Shulin Tian, Ziqi Huang, Fan Zhang, Hongyuan Zhu, Yu Qiao, Ziwei Liu
• Venue: arXiv: cs.AI
• Published: 2026-08-10
• Collected: 2026-08-11
• URL: http://arxiv.org/abs/2608.09666v1
• PDF: https://arxiv.org/pdf/2608.09666v1
• DOI: UNKNOWN
• Registry group: academic_apis_and_preprints
• Evidence role: canonical_academic_index
• Language: en
• Country / region: UNKNOWN

Plain-Language Brief

Researchers present the Evaluation Agent framework, a system that efficiently evaluates visual generative models using human-like strategies. It decomposes evaluation requests into sub-aspects, generates prompts, samples images/videos, and iteratively updates its plan based on observed evidence. Experiments show it reduces evaluation time to 10% of traditional methods while delivering comparable results.

What the Document Says

• The paper introduces the Evaluation Agent framework for efficient evaluation of visual generative models.
• It uses human-like strategies for multi-round evaluations, providing detailed and user-tailored analyses.
• Given a natural-language evaluation request, it decomposes it into sub-aspects, generates targeted prompts, samples images/videos, invokes suitable evaluation tools, and iteratively updates its plan.
• Experiments show that the Evaluation Agent reduces evaluation time to 10% of traditional methods while delivering comparable results.
• The authors also introduce Open Evaluation Agent (Open-EA) by constructing EA-CoT-10K and training EA-3B as a local planning backbone.

Key Concepts Mentioned

• Evaluation Agent
• Visual Generative Models
• Efficient evaluation
• Multi-round evaluations
• API-based agent
• Open Evaluation Agent (Open-EA)
• EA-CoT-10K
• EA-3B

Claims Made by the Document

• The Evaluation Agent can efficiently evaluate visual generative models.
• Open Evaluation Agent reduces dependence on proprietary backbones.

Evidence Used

• The paper describes the framework and its strategies for efficient evaluation.
• Experiments showing reduced evaluation time are presented.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Recent advances in visual generative models have enabled high-quality image and video generation, but evaluating these models often demands sampling hundreds or thousands of images or videos, which is computationally expensive. Existing evaluation methods also rely on rigid pipelines that overlook specific user needs and provide numerical results without clear explanations. Mimicking how humans quickly form impressions of a model's capabilities from only a few samples, we propose the Evaluation Agent framework, which employs human-like strategies for efficient, dynamic, multi-round evaluations, offering detailed, user-tailored analyses. Given a natural-language evaluation request, the agent decomposes it into sub-aspects, generates targeted prompts, samples images or videos from the evaluated model, invokes suitable evaluation tools, and iteratively updates its plan from the observed evidence, covering both predefined benchmark dimensions and open-ended user concerns. The framework is thus efficient, promptable, explainable, and scalable across models and tools. Experiments show that Evaluation Agent reduces evaluation time to 10% of traditional methods while delivering comparable results. We further introduce Open Evaluation Agent (Open-EA) by constructing EA-CoT-10K, a corpus of history-conditioned step-level instruction-tuning records derived from multi-round evaluation rollouts, and training EA-3B from Qwen2.5-3B-Instruct as a local planning backbone that preserves the structured reasoning, tool invocation, and summary protocol of the API-based agent while reducing dependence on proprietary backbones. Experiments validate the API-based agent on established T2I/T2V benchmarks and open-ended queries, and evaluate Open-EA on four in-domain and three out-of-domain T2V generator families, showing partial cross-family transfer of the learned policy.