Image · Text to Image · August 5, 2026
Text to Image Leaderboard
Every major text-to-image model ranked by Elo from blind head-to-head human votes — no cherry-picked grids, no lab-supplied samples.
Models ranked
45
Open weights
10
Arena appearances
418,384
Prompt categories
5
Highlights
Higher is better
Glsrm Arena
Overall leaderboard · best model per lab
- 1GPT Image 2 (high)1339
- 2Reve 2.11299
- 3MAI-Image-2.51270
- 4Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)1263
- 5Riverflow 2.01254
- 6HiDream-O1-Image-1.51245
- 7Seedream 5.0 Pro1240
- 8Cosmos3-Super-Text2Image (agentic)1219
- 9Recraft V4.1 Utility1205
Higher is better
Open Weights
Downloadable checkpoints · best per lab
- 1Cosmos3-Super-Text2Image (agentic)1219
- 2HiDream-O1-Image-Dev-26041188
- 3Ideogram 4.0 (Quality)1174
- 4ERNIE Image1166
- 5Ideogram 4.0 Fast (Quality)1159
- 6Qwen Image Max 25121158
- 7FLUX.2 [dev]1154
Lower is better
Price per Image
USD per generation, default quality · cheapest per lab
- 1FLUX.2 [dev] Turbo$0.01
- 2ERNIE Image Turbo$0.01
- 3FLUX.2 [dev]$0.01
- 4Krea 2 Medium Turbo$0.01
- 5MAI-Image-2.5-Flash$0.02
- 6grok-imagine-image$0.02
- 7Qwen Image Max 2512$0.02
- 8Reve 2.0$0.02
- 9Seedream 4.0$0.03
Global leaderboard
Elo from blind head-to-head votes · All categories
| Rank | Model | Creator | Elo | 95% CI | Appearances | $/Image | Released |
|---|---|---|---|---|---|---|---|
| 1 | GPT Image 2 (high) | OpenAI | 1339 | ±7 | 21,321 | $0.21 | Apr 2026 |
| 2 | Reve 2.1 | Reve | 1299 | ±11 | 9,889 | $0.20 | Jul 2026 |
| 3 | MAI-Image-2.5 | Microsoft AI | 1270 | ±9 | 9,868 | $0.05 | Jun 2026 |
| 4 | Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) | 1263 | ±9 | 9,900 | $0.03 | Jun 2026 | |
| 5 | GPT Image 1.5 (high) | OpenAI | 1263 | ±7 | 12,752 | $0.13 | Dec 2025 |
| 6 | Nano Banana 2 (Gemini 3.1 Flash Image Preview) | 1262 | ±7 | 19,616 | $0.07 | Feb 2026 | |
| 7 | Riverflow 2.0 | Sourceful | 1254 | ±8 | 7,637 | $0.15 | Feb 2026 |
| 8 | HiDream-O1-Image-1.5 | HiDream | 1245 | ±8 | 12,456 | $0.08 | Jun 2026 |
| 9 | Seedream 5.0 Pro | Bytedance | 1240 | ±9 | 8,271 | $0.09 | Jul 2026 |
| 10 | Reve 2.0 | Reve | 1238 | ±10 | 3,825 | $0.02 | Jun 2026 |
| 11 | Nano Banana Pro (Gemini 3 Pro Image) | 1225 | ±7 | 12,019 | $0.13 | Nov 2025 | |
| 12 | Cosmos3-Super-Text2Image (agentic)Open | NVIDIA | 1219 | ±8 | 10,414 | — | May 2026 |
| 13 | MAI-Image-2.5-Flash | Microsoft AI | 1208 | ±9 | 4,979 | $0.02 | Jun 2026 |
| 14 | Recraft V4.1 Utility | Recraft | 1205 | ±8 | 5,511 | $0.04 | May 2026 |
| 15 | grok-imagine-image-quality | xAI | 1204 | ±7 | 15,871 | $0.05 | Apr 2026 |
| 16 | Recraft V4.1 Utility Pro | Recraft | 1201 | ±7 | 15,341 | $0.21 | May 2026 |
| 17 | FLUX.2 [max] | Black Forest Labs | 1196 | ±7 | 11,711 | $0.07 | Dec 2025 |
| 18 | Seedream 4.0 | ByteDance Seed | 1192 | ±7 | 16,966 | $0.03 | Sep 2025 |
| 19 | Krea 2 Medium | Krea | 1191 | ±7 | 15,187 | $0.03 | May 2026 |
| 20 | Krea 2 Large | Krea | 1190 | ±7 | 15,547 | $0.06 | May 2026 |
| 21 | MAI-Image-2 | Microsoft AI | 1189 | ±7 | 7,744 | $0.04 | Mar 2026 |
| 22 | HiDream-O1-Image-Dev-2604Open | HiDream | 1188 | ±7 | 7,175 | — | May 2026 |
| 23 | Krea 2 Medium Turbo | Krea | 1188 | ±10 | 3,117 | $0.01 | Jun 2026 |
| 24 | FLUX.2 [pro] | Black Forest Labs | 1187 | ±7 | 11,026 | $0.03 | Nov 2025 |
| 25 | grok-imagine-image | xAI | 1181 | ±7 | 9,527 | $0.02 | Jan 2026 |
| 26 | FLUX.2 [flex] | Black Forest Labs | 1180 | ±8 | 6,427 | $0.06 | Nov 2025 |
| 27 | Cosmos3-Super-Text2Image-4StepOpen | NVIDIA | 1176 | ±10 | 6,402 | — | Jul 2026 |
| 28 | Luma UNI 1 Max | Luma Labs | 1175 | ±8 | 5,785 | $0.10 | May 2026 |
| 29 | Ideogram 4.0 (Quality)Open | Ideogram | 1174 | ±8 | 9,978 | $0.10 | Jun 2026 |
| 30 | Qwen Image 2.0 Pro (2026-04-22) | Alibaba | 1174 | ±8 | 5,820 | $0.07 | Apr 2026 |
| 31 | Imagen 4 Ultra | 1172 | ±8 | 6,804 | $0.06 | Jun 2025 | |
| 32 | Seedream 4.5 | ByteDance Seed | 1169 | ±7 | 12,711 | $0.04 | Dec 2025 |
| 33 | ImagineArt 2.0 | ImagineArt | 1169 | ±7 | 10,884 | $0.03 | Apr 2026 |
| 34 | ERNIE ImageOpen | Baidu | 1166 | ±8 | 4,813 | $0.03 | Apr 2026 |
| 35 | ERNIE Image TurboOpen | Baidu | 1163 | ±8 | 5,723 | $0.01 | Apr 2026 |
| 36 | MAI-Image-2-Efficient | Microsoft AI | 1159 | ±7 | 8,344 | $0.02 | Apr 2026 |
| 37 | Ideogram 4.0 Fast (Quality)Open | Fal | 1159 | ±10 | 3,699 | $0.02 | Jul 2026 |
| 38 | Qwen Image Max 2512Open | Alibaba | 1158 | ±8 | 5,727 | $0.02 | Dec 2025 |
| 39 | Recraft V4.1 Pro | Recraft | 1157 | ±8 | 5,115 | $0.21 | May 2026 |
| 40 | FLUX.2 [dev]Open | Black Forest Labs | 1154 | ±7 | 9,916 | $0.01 | Nov 2025 |
| 41 | Ideogram 4.0 | Ideogram | 1154 | ±10 | 3,461 | $0.06 | Jun 2026 |
| 42 | FLUX.2 [dev] TurboOpen | Fal | 1154 | ±8 | 6,348 | $0.01 | Dec 2025 |
| 43 | Recraft V4.1 | Recraft | 1153 | ±8 | 5,406 | $0.04 | May 2026 |
| 44 | Nano Banana (Gemini 2.5 Flash Image) | 1152 | ±7 | 11,147 | $0.04 | Aug 2025 | |
| 45 | Luma UNI 1 | Luma Labs | 1150 | ±8 | 6,204 | $0.04 | May 2026 |
Showing 45 of 45 models. Category Elo is computed from votes on that prompt type alone.
Glsrm Arena — Open Weights
Every downloadable checkpoint on the board, ranked by overall Elo. Higher is better.
Cosmos3 sits fifth overall — ahead of most closed APIs — and the FLUX.2 [dev] line keeps Black Forest Labs in four of the fourteen open slots.
Glsrm Arena vs. Price per Image
Crowd-judged quality against USD per generation (log scale).
Up and to the left wins. The open-weights mid-table — FLUX.2 [dev], Qwen-Image-2512, Hunyuan — is where the value lives; models without public per-image pricing are excluded.
Glsrm Arena Over Time
Glsrm Arena by release date. The dashed line tracks the running frontier.
The frontier has changed hands four times since Imagen 4 Ultra — Google, ByteDance Seed, and OpenAI have all held it. GPT Image 2 set the current mark in April 2026.
How the arena works
1 · Prompt
Prompts are sampled across five categories — people, photorealism, art, product, and typography — so no model can farm a single style.
2 · Blind matchup
Two models generate from the same prompt. Voters see the images side by side with no names, no logos, and no way to tell who made what.
3 · Elo
Each vote updates both models' Elo ratings. Confidence intervals come from resampling the vote history; new models stay provisional until they clear roughly 3,000 appearances.
FAQ
Which model leads the arena right now?
GPT Image 2 (high) holds the top spot with an Elo of 1339 across 21,321 appearances — a 40-point gap over the chasing pack.
And among open weights?
NVIDIA's Cosmos3-Super-Text2Image (agentic) leads the open-weights cut at 1219 — twelfth overall, ahead of plenty of closed APIs — followed by HiDream-O1-Image-Dev-2604 (1188) and NVIDIA's Cosmos3-Super-Text2Image-4Step (1176).
How are models ranked?
Pure crowd judgment. Voters get a prompt and two unlabeled images from different models and pick the better one; thousands of those blind matchups feed an Elo rating, with a 95% confidence interval from resampling the vote history.
What counts as open weights?
The checkpoint has to be downloadable and runnable on your own hardware under a published license. API-only siblings of open families — FLUX.2 [pro], Qwen Image 2.0 Pro — stay in the proprietary column.
Latest Research
Reporting from the paper mines
PolyU Unveils AI Virtual Patient for Cancer Care
By Mirage News · 2 hrs ago
Media Briefing: Publishers are replacing Google with a little bit of everything
By Digiday · 3 hrs ago
From deepfakes to rape threats: AI is supercharging abuse of women in public life
By Australian Broadcasting… · 6 hrs ago
University of Nottingham AI spin-out lands funding after showing proof of concept
By Insider Media Ltd · 2 hrs ago
Methodology: rankings come from blind pairwise human votes; Elo, confidence intervals, and appearance counts reflect the full vote history at snapshot time. Arena data: public image-arena snapshot, August 5, 2026. Prices are list API prices per image at default quality and change frequently.