whatmodel.sh / image / DALL·E 2
DALL·E 2
↗ share
+ contribute
DALL·E 2 is a closed-weights image model from OpenAI, released April 2022. Licence: proprietary.
Catalogue identity notes
OpenAI API model dall-e-2. OpenAI's deprecations page lists dall-e-2 shut down on 2026-05-12 (announced 2025-11-14), so api is null; kept to anchor the fit.
Score within the LMArena text-to-image comparison set only.
$ how it compares
The image models ranked either side of DALL·E 2, out of 197 with a score, then more from OpenAI.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 158 |  Janus-Pro-1B | 36±4 | api $0.0005/image +local | ||
| 159 |  Show-o (show-o-demo-512) | 36±3 | local | ||
| 160 | DALL·E 2 | 36±3 | comparison only | ||
| 161 |  SSD-1B | 36±4 | local | ||
| 162 | JanusFlow-1.3B | 36±6 | comparison only | ||
| more from OpenAI | |||||
| 1 | GPT Image 2.5 Sunburst | 92±2 | api $0.013/image | ||
| 2 | GPT Image 2.5 Flare | 88±2 | api $0.013/image | ||
| 3 | GPT Image 2 | 85±1 | api $0.053/image | ||
$ what it's good at
text to image▁▂▇▇▆▅█▇▂▃▁▁ #139 of 157 * · 34% of its score · 2 tests
Beats Latent Diffusion text2img-large (LDM-KL-8-G, LAION-400M), VQ-Diffusion (ITHQ) here, though they rank higher overall. Trails CogView2, Versatile Diffusion v1.0, Dreamlike Photoreal 2.0 here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| ImagenHub text-guided generation (human overall) | 54.0% ✎ | ▄▄·█·······▄ beats 1/4 | 34% |
| HPS v2 benchmark (4-style average) | 27.0% ✎ | ▃··▃▅·▅·▅▅█▅ beats 8/14 | 0.3% |
other tests2 not in a task · 61% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| ImageReward human ranking of T2I models (#wins, real user prompts) | 390.0000 wins ✎ | ▄·········█▄ beats 3/3 | 43% |
| HEIM text-image alignment (aspect win rate, v1.1.0) | 97.1% ✎ | ▁····█▁·▁··▃ beats 9/9 | 18% |
$ where to run it
No verified API or download recorded yet · add one
$ how the score was worked out · ImageReward: Learning and Eva… 43%, ImagenHub: Standardizing the… 34%, HEIM: Holistic Evaluation of… 18% and 1 more
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 ImageReward: Learning and Evaluating Human Preferences for Text-to-Image Generation (arXiv 2304.05977v4) | 1 | 43% |
| ImagenHub: Standardizing the evaluation of conditional image generation models (arXiv v4) | 1 | 34% |
| HEIM: Holistic Evaluation of Text-to-Image Models (arXiv 2311.04287v1) | 1 | 18% |
 tgxs002/HPSv2 repository README, HPS v2 benchmark table | 1 | 0% |
| prior · pull toward the middle | 5% | |
| weighted average | = 36.1 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json