AI Paper Atlas

Research map / Benchmarks / MS-COCO 30K

MS-COCO 30K leaderboard

15 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on MS-COCO 30K. Best FID: 4.42 (PerCo (0.012 bpp), ICLR 2024).

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.

MS-COCO 30K · FID

15 papers, each with its best reported number for its own method; lower is better.

#FID ↓MethodPaperTopic
14.42PerCo (0.012 bpp)Towards image compression with perfect realism at ultra-low bitrates
ICLR 2024 · Table: conditional diffusion a
Image restoration
27.4LaVIT (LLaMA-7B)
LLaMA-7B
Unified Language-Vision Pretraining in LLM with Dynamic Discrete Visual Tokenization
ICLR 2024 · zero-shot generation table
Human and embodied modeling
37.65SynerGen-VL
InternLM2-1.8B
SynerGen-VL: Towards Synergistic Image Understanding and Generation with Vision Experts and Token Folding
CVPR 2025 · Table: Image generation result
Multimodal LLMs
48.39Harmon-1.5B
Qwen2.5-1.5B-Instruct + MAR-H
Harmonizing Visual Representations for Unified Multimodal Understanding and Generation
ICCV 2025 · Table 3
Human and embodied modeling
58.53Janus
DeepSeek-LLM 1.3B
Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation
CVPR 2025 · Table 3
Human and embodied modeling
610.18TTS w/ NEP
RLlamaGen
NEP: Autoregressive Image Editing via Next Editing Token Prediction
NeurIPS 2025 · NEP pre-training and TTC table
Editing and style transfer
710.4ITIT-H
ViT-H
Leveraging Unpaired Data for Vision-Language Generative Models via Cycle Consistency
ICLR 2024 · Comparison to Prior Work table
Text-to-image diffusion
811.48SAG+TCFG
SD v1.4
TCFG: Tangential Damping Classifier-free Guidance
CVPR 2025 · Table 3
Fast diffusion sampling
912.2Semantic Surgery
Stable Diffusion v1.4
Semantic Surgery: Zero-Shot Concept Erasure in Diffusion Models
NeurIPS 2025 · Table: Explicit content remova
Text-to-image diffusion
1013.4OmniFlowOmniFlow: Any-to-Any Generation with Multi-Modal Rectified Flows
CVPR 2025 · Table 1
Text-to-image diffusion
5 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.