GenEval

GenEval is a benchmark used to evaluate the performance of models in unified multimodal understanding and generation tasks.

Papers using GenEval (21)

IV-CoT: Implicit Visual Chain-of-Thought for Structure-Aware Text-to-Image Generation2026Rethinking UMM Visual Generation: Masked Modeling for Efficient Image-Only Pre-training2026Can AI Draw Science? A Benchmark for Evaluating Scientific Figure Generation by Text-to-Image and Multimodal Models2026Mural: Transferring LLM knowledge to image generation via Mixture-of-Transformers2026Prompt Reinjection: Alleviating Prompt Forgetting in Multimodal Diffusion Transformers2026Dynin-Omni: Omnimodal Unified Large Diffusion Language Model2026EvoTok: A Unified Image Tokenizer via Residual Latent Evolution for Visual Understanding and Generation2026Cheers: Decoupling Patch Details from Semantic Representations Enables Unified Multimodal Comprehension and Generation2026HYDRA: Unifying Multi-modal Generation and Understanding via Representation-Harmonized Tokenization2026Enhancing Alignment for Unified Multimodal Models via Semantically-Grounded Supervision2026M3: High-fidelity Text-to-Image Generation via Multi-Modal, Multi-Agent and Multi-Round Visual Reasoning2026GenAgent: Scaling Text-to-Image Generation via Agentic Multimodal Reasoning2026MammothModa2: A Unified AR-Diffusion Framework for Multimodal Understanding and Generation2025UniGame: Turning a Unified Multimodal Model Into Its Own Adversary2025Reconstruction Alignment Improves Unified Multimodal Models2025Lavida-O: Elastic Large Masked Diffusion Models for Unified Multimodal Understanding and Generation2025The Telephone Game: Evaluating Semantic Drift in Unified Models2025Ovis-U1 Technical Report2025Unigen: Enhanced Training & Test-time Strategies For Unified Multimodal Understanding And Generation2025OpenUni: A Simple Baseline for Unified Multimodal Understanding and Generation2025Harmonizing Visual Representations for Unified Multimodal Understanding and Generation2025