LongBanana: An Expert-Verified Benchmark for Long-Context Multi-Reference Image Synthesis
cs.CV, cs.CL
Submitted: 2026-11-04
Updated: 2026-11-04
Code: https://github.com/TIGER-AI-Lab/LongBanana
Project page: https://long-banana.com
Terminology
Sources
- Emu3.5: Native Multimodal Models are World Learners
- VideoScore2: Think before You Score in Generative Video Evaluation
- VGI-Bench: Probing Visual Intelligence in Video Generation Models
- ELLA: Equip Diffusion Models with LLM for Enhanced Semantic Alignment
- Counting on Consensus: Selecting the Right Inter-annotator Agreement Metric for NLP Annotation and Evaluation
- GEditBench v2: A Human-Aligned Benchmark for General Image Editing
- FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space
- Visual Prototype Conditioned Focal Region Generation for UAV-Based Object Detection
- Human-Aligned MLLM Judges for Fine-Grained Image Editing Evaluation: A Benchmark, Framework, and Analysis
- EditScore: Unlocking Online RL for Image Editing via High-Fidelity Reward Modeling
- MultiBanana: A Challenging Benchmark for Multi-Reference Text-to-Image Generation
- xDailyBench: Benchmarking LLMs on Professional Consultation for Real-Life Problems
- Dr. Claw: An AI Scientist Workspace for Vibe Research
- CREval: An Automated Interpretable Evaluation for Creative Image Manipulation under Complex Instructions
- VR-Thinker: Boosting Video Reward Models through Thinking-with-Image Reasoning
- Qwen-Image Technical Report
- OmniGen2: Towards Instruction-Aligned Multimodal Generation
- EditReward: A Human-Aligned Reward Model for Instruction-Guided Image Editing
- MICON-Bench: Benchmarking and Enhancing Multi-Image Context Image Generation in Unified Multimodal Models
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs