Arena
Judged head to head, not scored on a test
Benchmarks measure whether an answer is correct. The arena measures whether a person prefers it — two models get the same brief, a judge picks a winner, and Elo does the rest.
Models with Elo
117
Categories
11
Highest Elo
1,479
Claude Opus 5.5
Tournaments run
1,131,445
Overall
The standing order
Elo pooled across every category a model competed in, alongside its raw win rate.
Overall Elo
Higher is better · scale runs 1,325–1,500, not from zero
- Claude Opus 5.5Anthropic1,479
- Claude Sonnet 5.5Anthropic1,468
- GPT-6 AstraOpenAI1,468
- GPT-6.1 SolOpenAI1,464
- Claude Fable 5.1Anthropic1,393
- Kimi K3Moonshot AI1,388
- Gemini 4 ArgonGoogle1,387
- GPT-5.6 SolOpenAI1,385
- Muse Spark 1.3Meta1,360
- V4.1 FlashDeepSeek1,357
- Claude Opus 5Anthropic1,349
- Claude Haiku 5.5Anthropic1,347
- Claude Fable 5Anthropic1,338
- Muse Spark 1.2Meta1,337
- GPT-6 SolOpenAI1,326
Elo is relative: a 100-point gap means the higher model wins roughly two thirds of head-to-heads. Zero Elo is meaningless, so this is a dot plot on a stated scale rather than bars from a zero baseline.
View as table
| Model | Elo | Win rate |
|---|---|---|
| Claude Opus 5.5 (Anthropic) | 1479 | 72.1% |
| Claude Sonnet 5.5 (Anthropic) | 1468 | 71.5% |
| GPT-6 Astra (OpenAI) | 1468 | 75.3% |
| GPT-6.1 Sol (OpenAI) | 1464 | 71.4% |
| Claude Fable 5.1 (Anthropic) | 1393 | 67.7% |
| Kimi K3 (Moonshot AI) | 1388 | 67.3% |
| Gemini 4 Argon (Google) | 1387 | 64.0% |
| GPT-5.6 Sol (OpenAI) | 1385 | 63.0% |
| Muse Spark 1.3 (Meta) | 1360 | 58.9% |
| V4.1 Flash (DeepSeek) | 1357 | 58.2% |
| Claude Opus 5 (Anthropic) | 1349 | 60.7% |
| Claude Haiku 5.5 (Anthropic) | 1347 | 59.0% |
| Claude Fable 5 (Anthropic) | 1338 | 60.8% |
| Muse Spark 1.2 (Meta) | 1337 | 58.0% |
| GPT-6 Sol (OpenAI) | 1326 | 58.4% |
Win rate
Share of head-to-heads won · higher is better
- GPT-6 AstraOpenAI75.3%
- Claude Opus 5.5Anthropic72.1%
- Claude Sonnet 5.5Anthropic71.5%
- GPT-6.1 SolOpenAI71.4%
- Gemini 2.5 ProGoogle68.2%
- Gemini 3.1 Pro PreviewGoogle67.8%
- Claude Fable 5.1Anthropic67.7%
- Kimi K3Moonshot AI67.3%
- 65.8%
- 64.4%
- Gemini 4 ArgonGoogle64.0%
- GPT-5.4OpenAI63.6%
- GPT-5.6 SolOpenAI63.0%
- Claude Opus 4.6Anthropic61.2%
- M2.1MiniMax60.9%
Win rate and Elo disagree when a model faced an unusually strong or weak field. The tick at 50% is even odds — anything left of it loses more than it wins.
View as table
| Model | Win rate |
|---|---|
| GPT-6 Astra (OpenAI) | 75.3% |
| Claude Opus 5.5 (Anthropic) | 72.1% |
| Claude Sonnet 5.5 (Anthropic) | 71.5% |
| GPT-6.1 Sol (OpenAI) | 71.4% |
| Gemini 2.5 Pro (Google) | 68.2% |
| Gemini 3.1 Pro Preview (Google) | 67.8% |
| Claude Fable 5.1 (Anthropic) | 67.7% |
| Kimi K3 (Moonshot AI) | 67.3% |
| Nano Banana Pro (Gemini 3 Pro Image Preview) (Google) | 65.8% |
| Nano Banana 2 (Gemini 3.1 Flash Image Preview) (Google) | 64.4% |
| Gemini 4 Argon (Google) | 64.0% |
| GPT-5.4 (OpenAI) | 63.6% |
| GPT-5.6 Sol (OpenAI) | 63.0% |
| Claude Opus 4.6 (Anthropic) | 61.2% |
| M2.1 (MiniMax) | 60.9% |
Cross-section
Nobody wins everywhere
The same twelve models across every well-covered category. Read down a column to see which brief separates the field.
Elo by model and category
Darker means stronger · blank means the model did not compete
| Model | 3D scenes | Data visualisation | Full-stack apps | Game development | Mobile apps | SVG | UI components | Websites |
|---|---|---|---|---|---|---|---|---|
| Claude Opus 5.5Anthropic | 1,479 | 1,400 | — | 1,459 | 1,327 | 1,332 | 1,380 | 1,337 |
| Claude Sonnet 5.5Anthropic | 1,456 | — | — | 1,468 | — | — | 1,381 | 1,331 |
| GPT-6 AstraOpenAI | 1,468 | 1,334 | 1,357 | 1,450 | 1,313 | 1,430 | 1,396 | 1,329 |
| GPT-6.1 SolOpenAI | 1,464 | — | — | 1,380 | — | — | 1,344 | 1,322 |
| Claude Fable 5.1Anthropic | 1,393 | 1,345 | 1,326 | 1,384 | 1,265 | 1,300 | 1,340 | 1,314 |
| Kimi K3Moonshot AI | 1,388 | 1,351 | 1,315 | 1,364 | 1,264 | 1,299 | 1,346 | 1,336 |
| Gemini 4 ArgonGoogle | 1,387 | — | — | 1,369 | — | — | 1,307 | 1,324 |
| GPT-5.6 SolOpenAI | 1,385 | 1,323 | 1,262 | 1,373 | 1,248 | 1,315 | 1,337 | 1,322 |
| Muse Spark 1.3Meta | 1,360 | 1,344 | 1,295 | 1,332 | 1,219 | 1,301 | 1,354 | 1,358 |
| V4.1 FlashDeepSeek | 1,357 | 1,293 | — | 1,333 | — | 1,243 | 1,318 | 1,311 |
| Claude Opus 5Anthropic | 1,335 | 1,329 | 1,306 | 1,340 | — | 1,310 | 1,349 | 1,310 |
| Claude Haiku 5.5Anthropic | 1,347 | 1,294 | — | 1,316 | — | — | 1,299 | 1,265 |
Only categories with at least eight competing models are shown, so a single lucky matchup cannot define a column.
View as table
| Model | 3D scenes | Data visualisation | Full-stack apps | Game development | Mobile apps | SVG | UI components | Websites |
|---|---|---|---|---|---|---|---|---|
| Claude Opus 5.5 | 1479 | 1400 | — | 1459 | 1327 | 1332 | 1380 | 1337 |
| Claude Sonnet 5.5 | 1456 | — | — | 1468 | — | — | 1381 | 1331 |
| GPT-6 Astra | 1468 | 1334 | 1357 | 1450 | 1313 | 1430 | 1396 | 1329 |
| GPT-6.1 Sol | 1464 | — | — | 1380 | — | — | 1344 | 1322 |
| Claude Fable 5.1 | 1393 | 1345 | 1326 | 1384 | 1265 | 1300 | 1340 | 1314 |
| Kimi K3 | 1388 | 1351 | 1315 | 1364 | 1264 | 1299 | 1346 | 1336 |
| Gemini 4 Argon | 1387 | — | — | 1369 | — | — | 1307 | 1324 |
| GPT-5.6 Sol | 1385 | 1323 | 1262 | 1373 | 1248 | 1315 | 1337 | 1322 |
| Muse Spark 1.3 | 1360 | 1344 | 1295 | 1332 | 1219 | 1301 | 1354 | 1358 |
| V4.1 Flash | 1357 | 1293 | — | 1333 | — | 1243 | 1318 | 1311 |
| Claude Opus 5 | 1335 | 1329 | 1306 | 1340 | — | 1310 | 1349 | 1310 |
| Claude Haiku 5.5 | 1347 | 1294 | — | 1316 | — | — | 1299 | 1265 |
By category
One board per brief
Data visualisation rewards different things than game development. These are the boards that matter if you know what you are building.
3D scenes
Arena Elo · dot plot on a stated scale, not bars from zero
- Claude Opus 5.5Anthropic1,479
- GPT-6 AstraOpenAI1,468
- GPT-6.1 SolOpenAI1,464
- Claude Sonnet 5.5Anthropic1,456
- Claude Fable 5.1Anthropic1,393
- Kimi K3Moonshot AI1,388
- Gemini 4 ArgonGoogle1,387
- GPT-5.6 SolOpenAI1,385
- Muse Spark 1.3Meta1,360
- V4.1 FlashDeepSeek1,357
View as table
| Model | Elo | Win rate |
|---|---|---|
| Claude Opus 5.5 | 1479 | 72.1% |
| GPT-6 Astra | 1468 | 75.3% |
| GPT-6.1 Sol | 1464 | 71.4% |
| Claude Sonnet 5.5 | 1456 | 69.4% |
| Claude Fable 5.1 | 1393 | 67.7% |
| Kimi K3 | 1388 | 67.3% |
| Gemini 4 Argon | 1387 | 64.0% |
| GPT-5.6 Sol | 1385 | 63.0% |
| Muse Spark 1.3 | 1360 | 58.9% |
| V4.1 Flash | 1357 | 58.2% |
Data visualisation
Arena Elo · dot plot on a stated scale, not bars from zero
- Claude Opus 5.5Anthropic1,400
- Kimi K3Moonshot AI1,351
- Claude Fable 5.1Anthropic1,345
- Muse Spark 1.3Meta1,344
- Muse Spark 1.2Meta1,337
- GPT-6 AstraOpenAI1,334
- Claude Opus 5Anthropic1,329
- GPT-5.6 SolOpenAI1,323
- Gemini 3.7 FlashGoogle1,317
- Claude Fable 5Anthropic1,305
View as table
| Model | Elo | Win rate |
|---|---|---|
| Claude Opus 5.5 | 1400 | 64.8% |
| Kimi K3 | 1351 | 63.2% |
| Claude Fable 5.1 | 1345 | 61.8% |
| Muse Spark 1.3 | 1344 | 58.1% |
| Muse Spark 1.2 | 1337 | 58.0% |
| GPT-6 Astra | 1334 | 60.4% |
| Claude Opus 5 | 1329 | 58.2% |
| GPT-5.6 Sol | 1323 | 55.6% |
| Gemini 3.7 Flash | 1317 | 55.3% |
| Claude Fable 5 | 1305 | 57.5% |
Full-stack apps
Arena Elo · dot plot on a stated scale, not bars from zero
- GPT-6 AstraOpenAI1,357
- Claude Fable 5.1Anthropic1,326
- Kimi K3Moonshot AI1,315
- Claude Opus 5Anthropic1,306
- 1,305
- Muse Spark 1.3Meta1,295
- Claude Fable 5Anthropic1,280
- GPT-5.6 SolOpenAI1,262
- Claude Sonnet 5Anthropic1,249
- Claude Opus 4.8Anthropic1,249
View as table
| Model | Elo | Win rate |
|---|---|---|
| GPT-6 Astra | 1357 | 61.1% |
| Claude Fable 5.1 | 1326 | 57.2% |
| Kimi K3 | 1315 | 61.8% |
| Claude Opus 5 | 1306 | 57.3% |
| Qwen3.8 Max (0902) | 1305 | 62.0% |
| Muse Spark 1.3 | 1295 | 56.2% |
| Claude Fable 5 | 1280 | 58.2% |
| GPT-5.6 Sol | 1262 | 54.3% |
| Claude Sonnet 5 | 1249 | 53.7% |
| Claude Opus 4.8 | 1249 | 55.3% |
Game development
Arena Elo · dot plot on a stated scale, not bars from zero
- Claude Sonnet 5.5Anthropic1,468
- Claude Opus 5.5Anthropic1,459
- GPT-6 AstraOpenAI1,450
- Claude Fable 5.1Anthropic1,384
- GPT-6.1 SolOpenAI1,380
- GPT-5.6 SolOpenAI1,373
- Gemini 4 ArgonGoogle1,369
- Kimi K3Moonshot AI1,364
- Claude Opus 5Anthropic1,340
- Claude Fable 5Anthropic1,338
View as table
| Model | Elo | Win rate |
|---|---|---|
| Claude Sonnet 5.5 | 1468 | 71.5% |
| Claude Opus 5.5 | 1459 | 71.8% |
| GPT-6 Astra | 1450 | 72.7% |
| Claude Fable 5.1 | 1384 | 65.6% |
| GPT-6.1 Sol | 1380 | 61.9% |
| GPT-5.6 Sol | 1373 | 62.1% |
| Gemini 4 Argon | 1369 | 62.0% |
| Kimi K3 | 1364 | 60.9% |
| Claude Opus 5 | 1340 | 58.4% |
| Claude Fable 5 | 1338 | 60.8% |
Graphic design
Arena Elo · dot plot on a stated scale, not bars from zero
- 1,249
- 1,249
- 1,226
- 1,151
View as table
| Model | Elo | Win rate |
|---|---|---|
| Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) | 1249 | 56.7% |
| Nano Banana 2 (Gemini 3.1 Flash Image Preview) | 1249 | 65.0% |
| Nano Banana Pro (Gemini 3 Pro Image Preview) | 1226 | 65.8% |
| Nano Banana (Gemini 2.5 Flash Image) | 1151 | 52.5% |
Image
Arena Elo · dot plot on a stated scale, not bars from zero
- 1,269
- 1,239
- 1,216
- 1,179
View as table
| Model | Elo | Win rate |
|---|---|---|
| Nano Banana 2 (Gemini 3.1 Flash Image Preview) | 1269 | 64.4% |
| Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) | 1239 | 53.3% |
| Nano Banana Pro (Gemini 3 Pro Image Preview) | 1216 | 60.4% |
| Nano Banana (Gemini 2.5 Flash Image) | 1179 | 53.4% |
logo
Arena Elo · dot plot on a stated scale, not bars from zero
- 1,248
- 1,214
- 1,181
- 1,169
View as table
| Model | Elo | Win rate |
|---|---|---|
| Nano Banana 2 (Gemini 3.1 Flash Image Preview) | 1248 | 61.1% |
| Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) | 1214 | 49.0% |
| Nano Banana Pro (Gemini 3 Pro Image Preview) | 1181 | 54.2% |
| Nano Banana (Gemini 2.5 Flash Image) | 1169 | 51.4% |
Mobile apps
Arena Elo · dot plot on a stated scale, not bars from zero
- Claude Opus 5.5Anthropic1,327
- GPT-6 AstraOpenAI1,313
- Claude Fable 5.1Anthropic1,265
- Kimi K3Moonshot AI1,264
- Claude Fable 5Anthropic1,260
- GPT-5.6 SolOpenAI1,248
- Gemini 3.8 FlashGoogle1,233
- Gemini 3.7 FlashGoogle1,227
- Claude Opus 4.8Anthropic1,220
- Muse Spark 1.3Meta1,219
View as table
| Model | Elo | Win rate |
|---|---|---|
| Claude Opus 5.5 | 1327 | 64.2% |
| GPT-6 Astra | 1313 | 65.0% |
| Claude Fable 5.1 | 1265 | 56.3% |
| Kimi K3 | 1264 | 57.0% |
| Claude Fable 5 | 1260 | 60.4% |
| GPT-5.6 Sol | 1248 | 54.5% |
| Gemini 3.8 Flash | 1233 | 54.4% |
| Gemini 3.7 Flash | 1227 | 52.9% |
| Claude Opus 4.8 | 1220 | 55.5% |
| Muse Spark 1.3 | 1219 | 47.9% |
SVG
Arena Elo · dot plot on a stated scale, not bars from zero
- GPT-6 AstraOpenAI1,430
- Claude Opus 5.5Anthropic1,332
- GPT-6 SolOpenAI1,326
- GPT-5.6 SolOpenAI1,315
- Claude Opus 5Anthropic1,310
- Muse Spark 1.3Meta1,301
- Claude Fable 5Anthropic1,301
- Claude Fable 5.1Anthropic1,300
- Kimi K3Moonshot AI1,299
- Gemini 3.8 FlashGoogle1,292
View as table
| Model | Elo | Win rate |
|---|---|---|
| GPT-6 Astra | 1430 | 68.6% |
| Claude Opus 5.5 | 1332 | 51.8% |
| GPT-6 Sol | 1326 | 58.4% |
| GPT-5.6 Sol | 1315 | 57.9% |
| Claude Opus 5 | 1310 | 58.4% |
| Muse Spark 1.3 | 1301 | 55.3% |
| Claude Fable 5 | 1301 | 62.6% |
| Claude Fable 5.1 | 1300 | 56.4% |
| Kimi K3 | 1299 | 59.2% |
| Gemini 3.8 Flash | 1292 | 54.7% |
UI components
Arena Elo · dot plot on a stated scale, not bars from zero
- GPT-6 AstraOpenAI1,396
- Claude Sonnet 5.5Anthropic1,381
- Claude Opus 5.5Anthropic1,380
- Muse Spark 1.3Meta1,354
- Claude Opus 5Anthropic1,349
- Kimi K3Moonshot AI1,346
- GPT-6.1 SolOpenAI1,344
- Claude Fable 5.1Anthropic1,340
- GPT-5.6 SolOpenAI1,337
- Gemini 3.8 FlashGoogle1,321
View as table
| Model | Elo | Win rate |
|---|---|---|
| GPT-6 Astra | 1396 | 65.6% |
| Claude Sonnet 5.5 | 1381 | 63.7% |
| Claude Opus 5.5 | 1380 | 63.3% |
| Muse Spark 1.3 | 1354 | 57.8% |
| Claude Opus 5 | 1349 | 60.7% |
| Kimi K3 | 1346 | 60.3% |
| GPT-6.1 Sol | 1344 | 57.6% |
| Claude Fable 5.1 | 1340 | 59.6% |
| GPT-5.6 Sol | 1337 | 58.1% |
| Gemini 3.8 Flash | 1321 | 53.9% |
Websites
Arena Elo · dot plot on a stated scale, not bars from zero
- Muse Spark 1.3Meta1,358
- Claude Opus 5.5Anthropic1,337
- Kimi K3Moonshot AI1,336
- Claude Sonnet 5.5Anthropic1,331
- GPT-6 AstraOpenAI1,329
- Gemini 4 ArgonGoogle1,324
- GPT-6.1 SolOpenAI1,322
- GPT-5.6 SolOpenAI1,322
- Muse Spark 1.2Meta1,315
- Claude Fable 5.1Anthropic1,314
View as table
| Model | Elo | Win rate |
|---|---|---|
| Muse Spark 1.3 | 1358 | 58.2% |
| Claude Opus 5.5 | 1337 | 57.5% |
| Kimi K3 | 1336 | 59.2% |
| Claude Sonnet 5.5 | 1331 | 56.3% |
| GPT-6 Astra | 1329 | 57.8% |
| Gemini 4 Argon | 1324 | 56.4% |
| GPT-6.1 Sol | 1322 | 55.8% |
| GPT-5.6 Sol | 1322 | 56.3% |
| Muse Spark 1.2 | 1315 | 52.1% |
| Claude Fable 5.1 | 1314 | 56.3% |