AI Paper Atlas

Research map / Benchmarks / DreamBench

DreamBench leaderboard

22 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on DreamBench. Best CLIP-T: 0.357 (ConceptPrism, CVPR 2026).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

DreamBench · CLIP-T

22 papers, each with its best reported number for its own method; higher is better.

#CLIP-T ↑MethodPaperTopic
10.357ConceptPrism
Stable Diffusion v2.1
ConceptPrism: Concept Disentanglement in Personalized Diffusion Models via Residual Token Optimization
CVPR 2026 · Table 1
Text-to-image diffusion
20.344Diptych Prompting
FLUX
Large-Scale Text-to-Image Model with Inpainting is a Zero-Shot Subject-Driven Image Generator
CVPR 2025 · Table 4
Text-to-image diffusion
30.329RPOSubject-driven Text-to-Image Generation via Preference-based Reinforcement Learning
NeurIPS 2024 · Ablation study on regularizati
Text-to-image diffusion
40.326X2Edit
FLUX.1
X2Edit: Revisiting Arbitrary-Instruction Image Editing Through Self-Constructed Data and Task-Aware Representation Learning
AAAI 2026 · Table 3
Editing and style transfer
50.326UniRealUniReal: Universal Image Generation and Editing via Learning Real-world Dynamics
CVPR 2025 · Table 3
Video generation
60.325EchoGen-2B
Infinity-2B
EchoGen: Generating Visual Echoes in Any Scene via Feed-Forward Subject-Driven Auto-Regressive Model
ICLR 2026 · Table 1
Text-to-image diffusion
70.3249DETEX
Stable Diffusion v1-4
Decoupled Textual Embeddings for Customized Image Generation
AAAI 2024 · Quantitative comparisons table
Text-to-image diffusion
80.318DCoAR
Lumina-mGPT-7B
DCoAR: Deep Concept Injection into Unified Autoregressive Models for Personalized Text-to-Image Generation
CVPR 2026 · Table 1
Text-to-image diffusion
90.315OmniGen
Phi-3
OmniGen: Unified Image Generation
CVPR 2025 · Table 3
Text-to-image diffusion
100.313MM-R1
Lumina-mGPT
MM-R1: Unleashing the Power of Unified Multimodal Large Language Models for Personalized Image Generation
AAAI 2026 · Table 1
Multimodal LLMs
12 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

DreamBench · CLIP-I

21 papers, each with its best reported number for its own method; higher is better.

#CLIP-I ↑MethodPaperTopic
10.882WeGen
LLaMA-2-7B-chat
WeGen: A Unified Model for Interactive Multimodal Generation as We Chat
CVPR 2025 · Table 2
Human and embodied modeling
20.852Ours (CoDi-2)
Llama-2-7b-chat-hf
CoDi-2: In-Context Interleaved and Interactive Any-to-Any Generation
CVPR 2024 · Table 2 (left)
Multimodal LLMs
30.847Kosmos-G* (single image input)
Kosmos-1 MLLM + SD v1.5 U-Net
Kosmos-G: Generating Images in Context with Multimodal Large Language Models
ICLR 2024 · Table 1
Multimodal LLMs
40.842MM-R1
Lumina-mGPT
MM-R1: Unleashing the Power of Unified Multimodal Large Language Models for Personalized Image Generation
AAAI 2026 · Table 1
Multimodal LLMs
50.841UNIC-Adapter
SD3
UNIC-Adapter: Unified Image-instruction Adapter with Multi-modal Transformer for Image Generation
CVPR 2025 · Table 2
Text-to-image diffusion
60.838PatchDPO
IP-Adapter-Plus
PatchDPO: Patch-level DPO for Finetuning-free Personalized Image Generation
CVPR 2025 · Table 2
Autoregressive and latent generation
70.835EchoGen-2B
Infinity-2B
EchoGen: Generating Visual Echoes in Any Scene via Feed-Forward Subject-Driven Auto-Regressive Model
ICLR 2026 · Table 1
Text-to-image diffusion
80.833RPO
SD
Subject-driven Text-to-Image Generation via Preference-based Reinforcement Learning
NeurIPS 2024 · Table 1
Text-to-image diffusion
90.827CAFE
Stable Diffusion 2
Customization Assistant for Text-to-Image Generation
CVPR 2024 · Table 1
Text-to-image diffusion
100.826X2Edit
FLUX.1
X2Edit: Revisiting Arbitrary-Instruction Image Editing Through Self-Constructed Data and Task-Aware Representation Learning
AAAI 2026 · Table 3
Editing and style transfer
11 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

DreamBench · DINO

18 papers, each with its best reported number for its own method; higher is better.

#DINO ↑MethodPaperTopic
10.823WeGen
LLaMA-2-7B-chat
WeGen: A Unified Model for Interactive Multimodal Generation as We Chat
CVPR 2025 · Table 2
Human and embodied modeling
20.822X2Edit
FLUX.1
X2Edit: Revisiting Arbitrary-Instruction Image Editing Through Self-Constructed Data and Task-Aware Representation Learning
AAAI 2026 · Table 3
Editing and style transfer
30.817X2IX2I: Seamless Integration of Multimodal Understanding into Diffusion Transformer via Attention Distillation
ICCV 2025 · Table 2
Text-to-image diffusion
40.816UNIC-Adapter
SD3
UNIC-Adapter: Unified Image-instruction Adapter with Multi-modal Transformer for Image Generation
CVPR 2025 · Table 2
Text-to-image diffusion
50.786MM-R1
Lumina-mGPT
MM-R1: Unleashing the Power of Unified Multimodal Large Language Models for Personalized Image Generation
AAAI 2026 · Table 1
Multimodal LLMs
60.755EchoGen-2B
Infinity-2B
EchoGen: Generating Visual Echoes in Any Scene via Feed-Forward Subject-Driven Auto-Regressive Model
ICLR 2026 · Table 1
Text-to-image diffusion
70.741SuTI (Imagen)
Imagen
Subject-driven Text-to-Image Generation via Apprenticeship Learning
NeurIPS 2023 · Table 1
Text-to-image diffusion
80.727PatchDPO
IP-Adapter-Plus
PatchDPO: Patch-level DPO for Finetuning-free Personalized Image Generation
CVPR 2025 · Table 2
Autoregressive and latent generation
90.723DCoAR
Lumina-mGPT-7B
DCoAR: Deep Concept Injection into Unified Autoregressive Models for Personalized Text-to-Image Generation
CVPR 2026 · Table 1
Text-to-image diffusion
100.715CAFE
Stable Diffusion 2
Customization Assistant for Text-to-Image Generation
CVPR 2024 · Table 1
Text-to-image diffusion
8 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.