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
100
Categories
11
Highest Elo
1,455
Kimi K3
Tournaments run
663,461
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 1,300–1,475, not from zero
- Kimi K3Moonshot AI1,455
- Claude Fable 5Anthropic1,397
- Claude Opus 5Anthropic1,393
- GPT-5.6 SolOpenAI1,380
- GLM 5.2Z.ai1,363
- Gemini 3.6 FlashGoogle1,346
- Claude Opus 4.7Anthropic1,337
- Claude Sonnet 5Anthropic1,336
- GPT-5.5OpenAI1,336
- Gemini 3.1 Pro PreviewGoogle1,333
- Kimi K2.6Moonshot AI1,329
- Claude Opus 4.6Anthropic1,329
- Muse Spark 1.2Meta1,322
- Qwen3.7 MaxQwen1,322
- MiMo-V2.5-ProXiaomi1,314
1,3001,475
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 |
|---|---|---|
| Kimi K3 (Moonshot AI) | 1455 | 69.4% |
| Claude Fable 5 (Anthropic) | 1397 | 65.4% |
| Claude Opus 5 (Anthropic) | 1393 | 65.2% |
| GPT-5.6 Sol (OpenAI) | 1380 | 62.7% |
| GLM 5.2 (Z.ai) | 1363 | 59.7% |
| Gemini 3.6 Flash (Google) | 1346 | 58.0% |
| Claude Opus 4.7 (Anthropic) | 1337 | 60.0% |
| Claude Sonnet 5 (Anthropic) | 1336 | 55.5% |
| GPT-5.5 (OpenAI) | 1336 | 60.7% |
| Gemini 3.1 Pro Preview (Google) | 1333 | 68.8% |
| Kimi K2.6 (Moonshot AI) | 1329 | 60.3% |
| Claude Opus 4.6 (Anthropic) | 1329 | 62.5% |
| Muse Spark 1.2 (Meta) | 1322 | 55.9% |
| Qwen3.7 Max (Qwen) | 1322 | 56.8% |
| MiMo-V2.5-Pro (Xiaomi) | 1314 | 58.7% |
Win rate
Share of head-to-heads won · higher is better
- Gemini 2.5 ProGoogle71.8%
- Kimi K3Moonshot AI69.4%
- Gemini 3.1 Pro PreviewGoogle68.8%
- 65.8%
- Claude Fable 5Anthropic65.4%
- Claude Opus 5Anthropic65.2%
- 64.4%
- GPT-5.4OpenAI63.6%
- GPT-5OpenAI62.9%
- GPT-5.6 SolOpenAI62.7%
- Gemini 3 Flash PreviewGoogle62.7%
- Claude Opus 4.6Anthropic62.5%
- M2.1MiniMax60.9%
- GPT-5.5OpenAI60.7%
- Kimi K2.6Moonshot AI60.3%
40even odds80
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 |
|---|---|
| Gemini 2.5 Pro (Google) | 71.8% |
| Kimi K3 (Moonshot AI) | 69.4% |
| Gemini 3.1 Pro Preview (Google) | 68.8% |
| Nano Banana Pro (Gemini 3 Pro Image Preview) (Google) | 65.8% |
| Claude Fable 5 (Anthropic) | 65.4% |
| Claude Opus 5 (Anthropic) | 65.2% |
| Nano Banana 2 (Gemini 3.1 Flash Image Preview) (Google) | 64.4% |
| GPT-5.4 (OpenAI) | 63.6% |
| GPT-5 (OpenAI) | 62.9% |
| GPT-5.6 Sol (OpenAI) | 62.7% |
| Gemini 3 Flash Preview (Google) | 62.7% |
| Claude Opus 4.6 (Anthropic) | 62.5% |
| M2.1 (MiniMax) | 60.9% |
| GPT-5.5 (OpenAI) | 60.7% |
| Kimi K2.6 (Moonshot AI) | 60.3% |
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 |
|---|---|---|---|---|---|---|---|---|
| Kimi K3Moonshot AI | 1,455 | 1,381 | — | — | — | — | 1,394 | 1,378 |
| Claude Fable 5Anthropic | 1,376 | 1,344 | 1,295 | 1,397 | 1,254 | 1,352 | 1,355 | 1,327 |
| Claude Opus 5Anthropic | 1,390 | 1,393 | — | — | — | — | 1,384 | 1,339 |
| GPT-5.6 SolOpenAI | 1,364 | 1,343 | — | 1,380 | — | 1,365 | 1,365 | 1,344 |
| GLM 5.2Z.ai | 1,363 | 1,325 | 1,274 | 1,337 | 1,219 | 1,260 | 1,339 | 1,337 |
| Gemini 3.6 FlashGoogle | 1,332 | 1,346 | — | 1,288 | — | — | 1,341 | 1,322 |
| Claude Opus 4.7Anthropic | 1,305 | 1,311 | — | 1,324 | — | 1,270 | 1,337 | 1,317 |
| Claude Sonnet 5Anthropic | 1,308 | 1,264 | 1,272 | 1,336 | 1,237 | 1,238 | 1,312 | 1,296 |
| GPT-5.5OpenAI | 1,250 | 1,283 | 1,128 | 1,336 | 1,209 | 1,278 | 1,292 | 1,280 |
| Gemini 3.1 Pro PreviewGoogle | 1,288 | 1,238 | 1,108 | 1,241 | 1,151 | 1,333 | 1,262 | 1,255 |
| Kimi K2.6Moonshot AI | 1,329 | 1,286 | 1,198 | 1,290 | 1,221 | 1,225 | 1,299 | 1,297 |
| Claude Opus 4.6Anthropic | 1,329 | 1,310 | 1,262 | 1,320 | 1,279 | 1,275 | 1,320 | 1,315 |
1,108High
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 |
|---|---|---|---|---|---|---|---|---|
| Kimi K3 | 1455 | 1381 | — | — | — | — | 1394 | 1378 |
| Claude Fable 5 | 1376 | 1344 | 1295 | 1397 | 1254 | 1352 | 1355 | 1327 |
| Claude Opus 5 | 1390 | 1393 | — | — | — | — | 1384 | 1339 |
| GPT-5.6 Sol | 1364 | 1343 | — | 1380 | — | 1365 | 1365 | 1344 |
| GLM 5.2 | 1363 | 1325 | 1274 | 1337 | 1219 | 1260 | 1339 | 1337 |
| Gemini 3.6 Flash | 1332 | 1346 | — | 1288 | — | — | 1341 | 1322 |
| Claude Opus 4.7 | 1305 | 1311 | — | 1324 | — | 1270 | 1337 | 1317 |
| Claude Sonnet 5 | 1308 | 1264 | 1272 | 1336 | 1237 | 1238 | 1312 | 1296 |
| GPT-5.5 | 1250 | 1283 | 1128 | 1336 | 1209 | 1278 | 1292 | 1280 |
| Gemini 3.1 Pro Preview | 1288 | 1238 | 1108 | 1241 | 1151 | 1333 | 1262 | 1255 |
| Kimi K2.6 | 1329 | 1286 | 1198 | 1290 | 1221 | 1225 | 1299 | 1297 |
| Claude Opus 4.6 | 1329 | 1310 | 1262 | 1320 | 1279 | 1275 | 1320 | 1315 |
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
- Kimi K3Moonshot AI1,455
- Claude Opus 5Anthropic1,390
- Claude Fable 5Anthropic1,376
- GPT-5.6 SolOpenAI1,364
- GLM 5.2Z.ai1,363
- Gemini 3.6 FlashGoogle1,332
- Kimi K2.6Moonshot AI1,329
- Claude Opus 4.6Anthropic1,329
- Qwen3.7 MaxQwen1,322
- V4 ProDeepSeek1,313
1,3001,475
View as table
| Model | Elo | Win rate |
|---|---|---|
| Kimi K3 | 1455 | 69.4% |
| Claude Opus 5 | 1390 | 64.7% |
| Claude Fable 5 | 1376 | 62.9% |
| GPT-5.6 Sol | 1364 | 59.3% |
| GLM 5.2 | 1363 | 59.7% |
| Gemini 3.6 Flash | 1332 | 54.2% |
| Kimi K2.6 | 1329 | 60.3% |
| Claude Opus 4.6 | 1329 | 62.5% |
| Qwen3.7 Max | 1322 | 56.8% |
| V4 Pro | 1313 | 58.6% |
Data visualisation
Arena Elo · dot plot on a stated scale, not bars from zero
- Claude Opus 5Anthropic1,393
- Kimi K3Moonshot AI1,381
- Gemini 3.6 FlashGoogle1,346
- Claude Fable 5Anthropic1,344
- GPT-5.6 SolOpenAI1,343
- GLM 5.2Z.ai1,325
- Qwen3.7 MaxQwen1,318
- Claude Opus 4.7Anthropic1,311
- Claude Opus 4.6Anthropic1,310
- Claude Sonnet 4.6Anthropic1,308
1,3001,400
View as table
| Model | Elo | Win rate |
|---|---|---|
| Claude Opus 5 | 1393 | 65.2% |
| Kimi K3 | 1381 | 65.4% |
| Gemini 3.6 Flash | 1346 | 58.0% |
| Claude Fable 5 | 1344 | 59.3% |
| GPT-5.6 Sol | 1343 | 57.9% |
| GLM 5.2 | 1325 | 57.1% |
| Qwen3.7 Max | 1318 | 58.1% |
| Claude Opus 4.7 | 1311 | 58.2% |
| Claude Opus 4.6 | 1310 | 58.7% |
| Claude Sonnet 4.6 | 1308 | 58.2% |
Full-stack apps
Arena Elo · dot plot on a stated scale, not bars from zero
- Claude Fable 5Anthropic1,295
- Claude Opus 4.8Anthropic1,285
- GLM 5.2Z.ai1,274
- Claude Sonnet 5Anthropic1,272
- Claude Opus 4.6Anthropic1,262
- Claude Sonnet 4.6Anthropic1,252
- Qwen3.7 MaxQwen1,240
- Gemini 3.5 FlashGoogle1,234
- MiniMax M3MiniMax1,226
- Kimi K2.7 CodeMoonshot AI1,218
1,2001,300
View as table
| Model | Elo | Win rate |
|---|---|---|
| Claude Fable 5 | 1295 | 63.7% |
| Claude Opus 4.8 | 1285 | 62.2% |
| GLM 5.2 | 1274 | 63.8% |
| Claude Sonnet 5 | 1272 | 59.8% |
| Claude Opus 4.6 | 1262 | 68.5% |
| Claude Sonnet 4.6 | 1252 | 64.1% |
| Qwen3.7 Max | 1240 | 55.6% |
| Gemini 3.5 Flash | 1234 | 56.7% |
| MiniMax M3 | 1226 | 53.0% |
| Kimi K2.7 Code | 1218 | 54.5% |
Game development
Arena Elo · dot plot on a stated scale, not bars from zero
- Claude Fable 5Anthropic1,397
- GPT-5.6 SolOpenAI1,380
- GLM 5.2Z.ai1,337
- Claude Sonnet 5Anthropic1,336
- GPT-5.5OpenAI1,336
- Claude Opus 4.7Anthropic1,324
- Muse Spark 1.2Meta1,322
- Claude Opus 4.6Anthropic1,320
- MiMo-V2.5-ProXiaomi1,314
- Qwen3.7 MaxQwen1,311
1,3001,400
View as table
| Model | Elo | Win rate |
|---|---|---|
| Claude Fable 5 | 1397 | 65.4% |
| GPT-5.6 Sol | 1380 | 62.7% |
| GLM 5.2 | 1337 | 59.1% |
| Claude Sonnet 5 | 1336 | 55.5% |
| GPT-5.5 | 1336 | 60.7% |
| Claude Opus 4.7 | 1324 | 60.0% |
| Muse Spark 1.2 | 1322 | 55.9% |
| Claude Opus 4.6 | 1320 | 61.2% |
| MiMo-V2.5-Pro | 1314 | 58.7% |
| Qwen3.7 Max | 1311 | 57.6% |
Graphic design
Arena Elo · dot plot on a stated scale, not bars from zero
- 1,277
- 1,268
- 1,254
- 1,179
1,1751,300
View as table
| Model | Elo | Win rate |
|---|---|---|
| Nano Banana 2 (Gemini 3.1 Flash Image Preview) | 1277 | 65.0% |
| Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) | 1268 | 55.5% |
| Nano Banana Pro (Gemini 3 Pro Image Preview) | 1254 | 65.8% |
| Nano Banana (Gemini 2.5 Flash Image) | 1179 | 52.5% |
Image
Arena Elo · dot plot on a stated scale, not bars from zero
- 1,291
- 1,260
- 1,239
- 1,201
1,2001,300
View as table
| Model | Elo | Win rate |
|---|---|---|
| Nano Banana 2 (Gemini 3.1 Flash Image Preview) | 1291 | 64.4% |
| Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) | 1260 | 53.9% |
| Nano Banana Pro (Gemini 3 Pro Image Preview) | 1239 | 60.4% |
| Nano Banana (Gemini 2.5 Flash Image) | 1201 | 53.4% |
logo
Arena Elo · dot plot on a stated scale, not bars from zero
- 1,262
- 1,224
- 1,195
- 1,183
1,1751,275
View as table
| Model | Elo | Win rate |
|---|---|---|
| Nano Banana 2 (Gemini 3.1 Flash Image Preview) | 1262 | 61.1% |
| Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) | 1224 | 48.5% |
| Nano Banana Pro (Gemini 3 Pro Image Preview) | 1195 | 54.2% |
| Nano Banana (Gemini 2.5 Flash Image) | 1183 | 51.4% |
Mobile apps
Arena Elo · dot plot on a stated scale, not bars from zero
- Claude Opus 4.6Anthropic1,279
- Claude Sonnet 4.6Anthropic1,271
- Claude Opus 4.8Anthropic1,255
- Claude Fable 5Anthropic1,254
- Claude Sonnet 5Anthropic1,237
- MiniMax M3MiniMax1,237
- Kimi K2.6Moonshot AI1,221
- GLM 5.2Z.ai1,219
- Gemini 3.5 FlashGoogle1,218
- GLM 5.1Z.ai1,218
1,2001,300
View as table
| Model | Elo | Win rate |
|---|---|---|
| Claude Opus 4.6 | 1279 | 64.8% |
| Claude Sonnet 4.6 | 1271 | 63.4% |
| Claude Opus 4.8 | 1255 | 57.5% |
| Claude Fable 5 | 1254 | 57.5% |
| Claude Sonnet 5 | 1237 | 55.1% |
| MiniMax M3 | 1237 | 55.4% |
| Kimi K2.6 | 1221 | 54.9% |
| GLM 5.2 | 1219 | 54.0% |
| Gemini 3.5 Flash | 1218 | 53.5% |
| GLM 5.1 | 1218 | 54.7% |
SVG
Arena Elo · dot plot on a stated scale, not bars from zero
- GPT-5.6 SolOpenAI1,365
- Claude Fable 5Anthropic1,352
- Gemini 3.1 Pro PreviewGoogle1,333
- Gemini 3.5 FlashGoogle1,296
- GPT-5.5OpenAI1,278
- Claude Opus 4.6Anthropic1,275
- Claude Opus 4.7Anthropic1,270
- Qwen3.7 MaxQwen1,265
- GLM 5.2Z.ai1,260
- GPT-5.6 TerraOpenAI1,254
1,2501,375
View as table
| Model | Elo | Win rate |
|---|---|---|
| GPT-5.6 Sol | 1365 | 64.7% |
| Claude Fable 5 | 1352 | 66.1% |
| Gemini 3.1 Pro Preview | 1333 | 68.8% |
| Gemini 3.5 Flash | 1296 | 60.9% |
| GPT-5.5 | 1278 | 58.1% |
| Claude Opus 4.6 | 1275 | 61.1% |
| Claude Opus 4.7 | 1270 | 59.2% |
| Qwen3.7 Max | 1265 | 59.4% |
| GLM 5.2 | 1260 | 55.5% |
| GPT-5.6 Terra | 1254 | 51.8% |
UI components
Arena Elo · dot plot on a stated scale, not bars from zero
- Kimi K3Moonshot AI1,394
- Claude Opus 5Anthropic1,384
- GPT-5.6 SolOpenAI1,365
- Claude Fable 5Anthropic1,355
- Gemini 3.6 FlashGoogle1,341
- GLM 5.2Z.ai1,339
- Claude Opus 4.7Anthropic1,337
- Claude Opus 4.6Anthropic1,320
- Claude Sonnet 5Anthropic1,312
- Qwen3.7 MaxQwen1,309
1,3001,400
View as table
| Model | Elo | Win rate |
|---|---|---|
| Kimi K3 | 1394 | 64.5% |
| Claude Opus 5 | 1384 | 62.0% |
| GPT-5.6 Sol | 1365 | 59.6% |
| Claude Fable 5 | 1355 | 59.1% |
| Gemini 3.6 Flash | 1341 | 55.9% |
| GLM 5.2 | 1339 | 58.3% |
| Claude Opus 4.7 | 1337 | 60.0% |
| Claude Opus 4.6 | 1320 | 60.0% |
| Claude Sonnet 5 | 1312 | 55.1% |
| Qwen3.7 Max | 1309 | 55.5% |
Websites
Arena Elo · dot plot on a stated scale, not bars from zero
- Kimi K3Moonshot AI1,378
- GPT-5.6 SolOpenAI1,344
- Claude Opus 5Anthropic1,339
- GLM 5.2Z.ai1,337
- Claude Fable 5Anthropic1,327
- Gemini 3.6 FlashGoogle1,322
- Muse Spark 1.2Meta1,318
- Claude Opus 4.7Anthropic1,317
- Claude Opus 4.6Anthropic1,315
- Claude Sonnet 4.6Anthropic1,308
1,3001,400
View as table
| Model | Elo | Win rate |
|---|---|---|
| Kimi K3 | 1378 | 63.5% |
| GPT-5.6 Sol | 1344 | 59.8% |
| Claude Opus 5 | 1339 | 58.9% |
| GLM 5.2 | 1337 | 60.1% |
| Claude Fable 5 | 1327 | 59.8% |
| Gemini 3.6 Flash | 1322 | 57.8% |
| Muse Spark 1.2 | 1318 | 55.7% |
| Claude Opus 4.7 | 1317 | 59.0% |
| Claude Opus 4.6 | 1315 | 61.4% |
| Claude Sonnet 4.6 | 1308 | 60.1% |