AI Paper Atlas

Research map / Benchmarks / DrawBench

DrawBench leaderboard

15 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on DrawBench. Best ImageReward: 1.71 (+ReNO +NoiseQuery, ICCV 2025).

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.

DrawBench · ImageReward

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

#ImageReward ↑MethodPaperTopic
11.71+ReNO +NoiseQuery
SD-Turbo
The Silent Assistant: NoiseQuery as Implicit Guidance for Goal-Driven Image Generation
ICCV 2025 · Table 1
Text-to-image diffusion
21.5959SOE
SDXL
Dual-Seed Evolutionary Algorithm for Noise Optimization in Diffusion Models
AAAI 2026 · Table 2
Text-to-image diffusion
31.319Ctrl-Z‡ (default)
Hunyuan-DiT
Ctrl-Z Sampling: Scaling Diffusion Sampling with Controlled Random Zigzag Explorations
ECCV 2026 · Table 1
Fast diffusion sampling
41.15RAISE (<=4 rounds)
FLUX.1-dev
RAISE: Requirement-Adaptive Evolutionary Refinement for Training-Free Text-to-Image Alignment
CVPR 2026 · Table 2
Text-to-image diffusion
51.15PixArt-Σ + Verifier Ensemble search
PixArt-Σ
Scaling Inference Time Compute for Diffusion Models
CVPR 2025 · Table 3
Fast diffusion sampling
61.08Ours (512 real-data pairs)
SD-3.5-M
When Preference Labels Fall Short: Aligning Diffusion Models from Real Data
ICML 2026 · Table 1a
Fast diffusion sampling
71.0284DDiT
FLUX-1.Dev
DDiT: Dynamic Patch Scheduling for Efficient Diffusion Transformers
CVPR 2026 · Table 1
Fast diffusion sampling
80.99GR-GRPO-7B
Janus-Pro-7B
GR-GRPO: Graph-Diffused Credit for Autoregressive Image RL Alignment
ECCV 2026 · Table 2
RL from feedback
90.847CogView3 (50+10 steps)CogView3: Finer and Faster Text-to-Image Generation via Relay Diffusion
ECCV 2024 · Table 1
Text-to-image diffusion
100.8305SCALE (Ours)
SDXL
SCALE: Semantic-Calibrated Guidance Enhancement for Prompt-Faithful Diffusion
ECCV 2026 · 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.