> For the complete documentation index, see [llms.txt](https://docs.cherryai.com.cn/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.cherryai.com.cn/docs/en-us/other/model_rank/text-to-image.md).
# Text-to-Image Rankings
This page shows [Arena AI](https://arena.ai/leaderboard/text-to-image) A daily snapshot of the top 30 on the leaderboard, for a quick view of recent model performance. For the full leaderboard, filters, and the latest changes, please visit the Arena AI website.
This leaderboard evaluates text-to-image models on their ability to generate images from text prompts.
> **Data updated at**: 2026-09-21 14:48:27 UTC / 2026-09-21 22:48:27 CST (Beijing Time)
{% hint style="info" %}
The leaderboard reflects specific evaluations and user voting preferences, and does not mean a model will definitely perform better on your task. When choosing a model, you should also consider price, speed, context length, tool calling, privacy, and regional availability.
{% endhint %}
## Top 30
| Rank | Rank range | Model | Score | Votes |
| ---: | ---------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ---------------------- | ------: |
| 1 | 1-2 | gpt-image-2.5-sunburst [1](https://openai.com/index/introducing-chatgpt-images-2-5/) | 1421 (±13) Preliminary | 3,149 |
| 2 | 1-2 | gpt-image-2.5-flare [1](https://openai.com/index/introducing-chatgpt-images-2-5/) | 1399 (±13) Preliminary | 2,856 |
| 3 | 3-3 | gpt-image-2 (medium) [1](https://x.com/OpenAI/status/2046665696898412887?s=20) | 1381 (±4) | 78,731 |
| 4 | 4-5 | mai-image-2.6 [1](https://microsoft.ai/news/mai-image-2-6-launches-at-no-2-on-arena-ahead-of-google-meta-and-xai/) | 1331 (±7) | 11,213 |
| 5 | 4-6 | grok-imagine-image-2.0 (low) [1](https://x.ai/news/grok-imagine-image-2) | 1315 (±12) Preliminary | 2,681 |
| 6 | 5-6 | reve-2.1 [1](http://reve.com/) | 1301 (±8) | 7,573 |
| 7 | 7-8 | muse-image [1](https://ai.meta.com/blog/introducing-muse-image-muse-video-msl/) | 1277 (±6) | 26,045 |
| 8 | 7-9 | reve-2.0 [1](http://reve.com/) | 1270 (±6) | 14,632 |
| 9 | 8-13 | gemini-3.1-flash-image (nano-banana-2) \[web-search] [1](http://aistudio.google.com/prompts/new_chat?model=gemini-3.1-flash-image) | 1261 (±5) | 41,957 |
| 10 | 9-13 | seedream-5.0-pro [1](https://seed.bytedance.com/en/seedream5_0_pro) | 1257 (±4) | 62,443 |
| 11 | 9-14 | qwen-image-3.0-pro [1](https://www.qwencloud.com/try-ai?scene=image\&models=qwen-image-3.0-pro) | 1254 (±7) | 10,920 |
| 12 | 9-13 | mai-image-2.5 [1](https://microsoft.ai/news/introducing-mai-image-2-5/) | 1254 (±4) | 57,250 |
| 13 | 9-14 | gemini-3.1-flash-lite-image (nano-banana-2-lite) [1](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-omni-flash-nano-banana-2-lite/) | 1250 (±6) | 16,371 |
| 14 | 12-14 | gemini-3-pro-image-2k (nano-banana-pro) [1](https://ai.studio/banana) | 1246 (±3) | 152,220 |
| 15 | 15-16 | gpt-image-1.5-high-fidelity [1](https://platform.openai.com/docs/models/gpt-image-1.5) | 1239 (±3) | 151,702 |
| 16 | 15-16 | gemini-3-pro-image-preview (nano-banana-pro) [1](https://ai.studio/banana) | 1232 (±5) | 82,737 |
| 17 | 17-17 | ideogram-4.0-quality [1](https://ideogram.ai/blog/4.0/) | 1204 (±4) | 42,169 |
| 18 | 18-21 | qwen-image-2.0-pro-2026-06-22 [1](https://bailian.console.alibabacloud.com/cn-beijing?tab=api#/api/?type=model\&url=2975126) | 1191 (±6) | 12,005 |
| 19 | 18-21 | uni-1.1-max [1](https://lumalabs.ai/api) | 1188 (±6) | 13,469 |
| 20 | 18-22 | mai-image-2 [1](https://microsoft.ai/news/introducing-MAI-Image-2) | 1183 (±5) | 49,198 |
| 21 | 18-22 | uni-1.1 [1](https://lumalabs.ai/api) | 1181 (±5) | 33,836 |
| 22 | 22-24 | grok-imagine-image [1](https://docs.x.ai/developers/model-capabilities/images/generation) | 1171 (±3) | 236,623 |
| 23 | 20-28 | recraft-v4.1-utility-pro [1](https://www.recraft.ai/blog/recraft-v4-1-more-beautiful-by-nature) | 1169 (±11) | 2,516 |
| 24 | 22-28 | Cosmos3-Super-Text2Image (Agentic) [1](https://huggingface.co/nvidia/Cosmos3-Super-Text2Image) | 1167 (±8) | 7,424 |
| 25 | 23-28 | flux-2-max [1](https://bfl.ai/models/flux-2) | 1162 (±4) | 117,339 |
| 26 | 23-29 | grok-imagine-image-pro [1](https://docs.x.ai/developers/model-capabilities/images/generation) | 1161 (±4) | 93,639 |
| 27 | 23-31 | flux-2-flex [1](https://bfl.ai/models/flux-2) | 1157 (±3) | 149,288 |
| 28 | 27-33 | flux-2-pro [1](https://bfl.ai/models/flux-2) | 1154 (±3) | 188,083 |
| 29 | 26-35 | reve-v1.5 [1](https://app.reve.com/) | 1154 (±4) | 36,748 |
| 30 | 23-35 | Cosmos3-Super-Text2Image [1](https://huggingface.co/nvidia/Cosmos3-Super-Text2Image) | 1154 (±7) | 12,499 |
## How to read this table
* **Rank / Rank range**: The relative ranking estimated by Arena AI based on text-to-image battle votes; the rank range indicates rank fluctuations within the confidence interval.
* **Score**: Relative score, suitable for comparing models on the leaderboard at the same time.
* **Votes**: Sample-size reference; when the sample size is small, rankings are usually more likely to fluctuate.
## Three things to confirm when choosing a model
1. Whether the provider actually offers this model, and whether it is available in your region and account;
2. Whether the API pricing, rate limits, and context length fit your task;
3. Run a small-scale test with 3–5 real tasks; don't rely only on the overall leaderboard ranking.
## Data source
Data from [Arena AI official text-to-image leaderboard](https://arena.ai/leaderboard/text-to-image), updated daily by GitHub Actions. For model pricing, licenses, and capabilities, please refer to the official information from the model provider.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://docs.cherryai.com.cn/docs/en-us/other/model_rank/text-to-image.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.