Live data from Arena AI
AI Benchmarks
Real ELO rankings from millions of human preference votes. 10 categories, updated weekly.
💬 Text / LLM 20 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | claude-fable-5 | Anthropic | 1508 | ±6 | 15,817 | proprietary |
| 🥈 | claude-opus-4-6-thinking | Anthropic | 1505 | ±4 | 64,435 | proprietary |
| 🥉 | claude-opus-4-7-thinking | Anthropic | 1502 | ±4 | 51,943 | proprietary |
| #4 | claude-opus-4-6 | Anthropic | 1498 | ±4 | 68,145 | proprietary |
| #5 | claude-opus-5-high | Anthropic | 1495 | ±8 | 5,417 | proprietary |
| #6 | claude-opus-4-7 | Anthropic | 1493 | ±4 | 53,093 | proprietary |
| #7 | muse-spark-1.1 | Meta | 1493 | ±7 | 9,573 | proprietary |
| #8 | muse-spark | Meta | 1488 | ±6 | 13,497 | proprietary |
| #9 | gemini-3.1-pro-preview | 1486 | ±3 | 86,298 | proprietary | |
| #10 | gemini-3-pro | 1486 | ±4 | 41,242 | proprietary | |
| #11 | kimi-k3 | Moonshot | 1485 | ±10 | 3,569 | proprietary |
| #12 | gpt-5.6-sol-xhigh | OpenAI | 1485 | ±7 | 7,972 | proprietary |
| #13 | claude-opus-4-8-thinking | Anthropic | 1484 | ±5 | 32,416 | proprietary |
| #14 | gemini-3.6-flash | 1482 | ±8 | 5,642 | proprietary | |
| #15 | gpt-5.5-high | OpenAI | 1482 | ±4 | 47,183 | proprietary |
| #16 | gpt-5.4-high | OpenAI | 1477 | ±4 | 60,260 | proprietary |
| #17 | gemini-3.5-flash-high | 1476 | ±7 | 10,016 | proprietary | |
| #18 | gpt-5.2-chat-latest-20260210 | OpenAI | 1476 | ±4 | 34,074 | proprietary |
| #19 | gpt-5.5 | OpenAI | 1476 | ±4 | 48,451 | proprietary |
| #20 | qwen3.7-max-preview | Alibaba | 1475 | ±10 | 3,698 | proprietary |
💻 Code 103 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | kimi-k3 | Moonshot | 1682 | ±13 | 3,776 | proprietary |
| 🥈 | claude-opus-5-high | Anthropic | 1673 | ±14 | 2,392 | proprietary |
| 🥉 | claude-fable-5 | Anthropic | 1629 | ±10 | 5,721 | proprietary |
| #4 | gpt-5.6-sol-xhigh (codex-harness) | OpenAI | 1625 | ±10 | 5,242 | proprietary |
| #5 | glm-5.2 (max) | Z.ai | 1587 | ±9 | 5,700 | open |
| #6 | claude-opus-4-8-thinking | Anthropic | 1568 | ±8 | 8,247 | proprietary |
| #7 | claude-opus-4-7 | Anthropic | 1559 | ±7 | 11,050 | proprietary |
| #8 | claude-opus-4-7-thinking | Anthropic | 1557 | ±7 | 11,661 | proprietary |
| #9 | grok-4.5 | SpaceXAI | 1549 | ±11 | 3,314 | proprietary |
| #10 | claude-opus-4-6-thinking | Anthropic | 1546 | ±6 | 13,607 | proprietary |
| #11 | claude-sonnet-5-high | Anthropic | 1543 | ±10 | 4,010 | proprietary |
| #12 | claude-opus-4-8 | Anthropic | 1539 | ±8 | 7,190 | proprietary |
| #13 | claude-opus-4-6 | Anthropic | 1538 | ±6 | 14,794 | proprietary |
| #14 | muse-spark-1.1 | Meta | 1536 | ±11 | 3,225 | proprietary |
| #15 | seed-2.1-pro-preview | Bytedance | 1529 | ±10 | 4,916 | proprietary |
| #16 | gemini-3.6-flash | 1524 | ±13 | 2,572 | proprietary | |
| #17 | claude-sonnet-4-6 | Anthropic | 1523 | ±6 | 16,764 | proprietary |
| #18 | glm-5.1 | Z.ai | 1518 | ±8 | 6,737 | open |
| #19 | hy3 | Tencent | 1518 | ±17 | 1,490 | open |
| #20 | qwen3.7-max-20260517 | Alibaba | 1517 | ±8 | 7,267 | proprietary |
| #21 | kimi-k2.6 | Moonshot | 1510 | ±8 | 9,263 | open |
| #22 | gpt-5.5-xhigh (codex-harness) | OpenAI | 1506 | ±7 | 10,168 | proprietary |
| #23 | claude-opus-4-5-20251101-thinking-32k | Anthropic | 1494 | ±8 | 10,539 | proprietary |
| #24 | minimax-m3 | MiniMax | 1493 | ±8 | 7,513 | open |
| #25 | gemini-3.5-flash | 1492 | ±14 | 2,313 | proprietary | |
| #26 | gpt-5.5-high (codex-harness) | OpenAI | 1485 | ±7 | 11,970 | proprietary |
| #27 | gemini-3.5-flash-medium | 1483 | ±9 | 6,646 | proprietary | |
| #28 | qwen3.6-max-preview | Alibaba | 1478 | ±13 | 2,546 | proprietary |
| #29 | mimo-v2.5-pro | Xiaomi | 1474 | ±7 | 10,686 | open |
| #30 | kimi-k2.7-code | Moonshot | 1473 | ±10 | 4,538 | open |
| #31 | claude-opus-4-5-20251101 | Anthropic | 1467 | ±7 | 12,782 | proprietary |
| #32 | deepseek-v4-pro-thinking | DeepSeek | 1464 | ±7 | 10,040 | open |
| #33 | gpt-5.4-high (codex-harness) | OpenAI | 1462 | ±20 | 1,261 | proprietary |
| #34 | qwen3.6-plus | Alibaba | 1458 | ±7 | 12,826 | proprietary |
| #35 | gpt-5.5 (codex-harness) | OpenAI | 1454 | ±7 | 10,313 | proprietary |
| #36 | gemini-3.5-flash-lite | 1447 | ±46 | 200 | proprietary | |
| #37 | deepseek-v4-pro | DeepSeek | 1447 | ±7 | 10,565 | open |
| #38 | gemini-3.1-pro-preview | 1445 | ±6 | 17,928 | proprietary | |
| #39 | gpt-5.4-medium (codex-harness) | OpenAI | 1444 | ±19 | 1,294 | proprietary |
| #40 | gemini-3-flash | 1438 | ±9 | 10,414 | proprietary | |
| #41 | gemini-3-pro | 1438 | ±9 | 13,485 | proprietary | |
| #42 | kimi-k2.5-thinking | Moonshot | 1436 | ±6 | 15,364 | open |
| #43 | mimo-v2.5 | Xiaomi | 1436 | ±7 | 9,637 | open |
| #44 | glm-5 | Z.ai | 1435 | ±8 | 6,967 | open |
| #45 | glm-4.7 | Z.ai | 1433 | ±12 | 3,796 | open |
| #46 | mimo-v2-pro | Xiaomi | 1433 | ±8 | 6,849 | proprietary |
| #47 | gpt-5.2 | OpenAI | 1419 | ±23 | 1,049 | proprietary |
| #48 | gpt-5-medium | OpenAI | 1419 | ±16 | 3,016 | proprietary |
| #49 | inkling | Thinky | 1419 | ±10 | 4,054 | open |
| #50 | gpt-5.3-codex (codex-harness) | OpenAI | 1409 | ±14 | 2,501 | proprietary |
| #51 | kimi-k2.5-instant | Moonshot | 1405 | ±12 | 3,087 | open |
| #52 | qwen3.5-397b-a17b | Alibaba | 1400 | ±6 | 15,127 | open |
| #53 | gpt-5.4-mini-high | OpenAI | 1399 | ±7 | 10,906 | proprietary |
| #54 | glm-5v-turbo | Z.ai | 1399 | ±14 | 2,071 | proprietary |
| #55 | minimax-m2.7 | MiniMax | 1398 | ±7 | 11,351 | open |
| #56 | claude-sonnet-4-5-20250929-thinking-32k | Anthropic | 1392 | ±8 | 12,734 | proprietary |
| #57 | gpt-5.1-medium | OpenAI | 1390 | ±12 | 4,737 | proprietary |
| #58 | gpt-5.4 | OpenAI | 1390 | ±19 | 1,132 | proprietary |
| #59 | claude-opus-4-1-20250805 | Anthropic | 1389 | ±11 | 6,735 | proprietary |
| #60 | minimax-m2.1-preview | MiniMax | 1388 | ±10 | 7,290 | open |
| #61 | minimax-m2.5 | MiniMax | 1386 | ±8 | 7,175 | open |
| #62 | claude-sonnet-4-5-20250929 | Anthropic | 1385 | ±7 | 15,080 | proprietary |
| #63 | gemini-3-flash (thinking-minimal) | 1383 | ±5 | 20,496 | proprietary | |
| #64 | grok-4.20-beta-0309-reasoning | SpaceXAI | 1374 | ±6 | 14,065 | proprietary |
| #65 | gpt-5.3-codex (codex-harness) | OpenAI | 1369 | ±12 | 3,425 | proprietary |
| #66 | gemma-4-26b-a4b | 1366 | ±17 | 1,417 | open | |
| #67 | gemma-4-31b | 1364 | ±8 | 6,994 | open | |
| #68 | deepseek-v3.2-thinking | DeepSeek | 1361 | ±9 | 6,613 | open |
| #69 | qwen3.5-122b-a10b | Alibaba | 1359 | ±8 | 7,694 | open |
| #70 | grok-4.3 | SpaceXAI | 1358 | ±7 | 9,480 | proprietary |
| #71 | qwen3.5-27b | Alibaba | 1357 | ±8 | 7,307 | open |
| #72 | hunyuan-hy3-preview | Tencent | 1357 | ±18 | 1,362 | |
| #73 | laguna-m.1 | Poolside | 1349 | ±10 | 4,860 | open |
| #74 | gpt-5.1 | OpenAI | 1341 | ±9 | 10,271 | proprietary |
| #75 | glm-4.6 | Z.ai | 1339 | ±11 | 6,568 | open |
| #76 | gpt-5.2-codex | OpenAI | 1338 | ±9 | 6,387 | proprietary |
| #77 | gpt-5.1-codex | OpenAI | 1336 | ±12 | 4,832 | proprietary |
| #78 | mimo-v2-flash (non-thinking) | Xiaomi | 1331 | ±10 | 5,411 | open |
| #79 | claude-haiku-4-5-20251001 | Anthropic | 1325 | ±5 | 23,498 | proprietary |
| #80 | deepseek-v3.2 | DeepSeek | 1323 | ±8 | 8,905 | open |
| #81 | kimi-k2-thinking-turbo | Moonshot | 1322 | ±7 | 12,547 | open |
| #82 | laguna-xs.2 | Poolside | 1303 | ±11 | 3,775 | open |
| #83 | minimax-m2 | MiniMax | 1297 | ±11 | 6,507 | open |
| #84 | mimo-v2-flash (thinking) | Xiaomi | 1291 | ±17 | 1,648 | open |
| #85 | qwen3-coder-480b-a35b-instruct | Alibaba | 1272 | ±8 | 12,219 | open |
| #86 | deepseek-v3.2-exp | DeepSeek | 1272 | ±14 | 3,845 | open |
| #87 | mistral-medium-3.5 | Mistral | 1267 | ±16 | 2,076 | open |
| #88 | gemini-3.1-flash-lite-preview | 1256 | ±7 | 13,328 | proprietary | |
| #89 | KAT-Coder-Pro-V1 | 1255 | ±20 | 1,449 | proprietary | |
| #90 | qwen3.5-35b-a3b | Alibaba | 1250 | ±18 | 1,521 | open |
| #91 | gpt-5.1-codex-mini | OpenAI | 1244 | ±22 | 1,127 | proprietary |
| #92 | grok-4-1-fast-reasoning | SpaceXAI | 1240 | ±11 | 5,309 | proprietary |
| #93 | trinity-large-thinking | 1239 | ±21 | 1,273 | open | |
| #94 | qwen3.5-flash | Alibaba | 1238 | ±20 | 1,256 | proprietary |
| #95 | mistral-large-3 | Mistral | 1230 | ±26 | 808 | open |
| #96 | gemini-2.5-pro | 1224 | ±16 | 2,657 | proprietary | |
| #97 | grok-4.1-thinking | SpaceXAI | 1210 | ±25 | 945 | proprietary |
| #98 | granite-4.1-8b | IBM | 1194 | ±19 | 1,684 | open |
| #99 | devstral-2 | Mistral | 1193 | ±21 | 1,136 | open |
| #100 | mercury-2 | Inception AI | 1166 | ±26 | 888 | proprietary |
| #101 | grok-code-fast-1 | SpaceXAI | 1163 | ±28 | 782 | proprietary |
| #102 | grok-4-fast-reasoning | SpaceXAI | 1160 | ±28 | 731 | proprietary |
| #103 | devstral-medium-2507 | Mistral | 1079 | ±31 | 825 | proprietary |
👁️ Vision 20 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | claude-fable-5 | Anthropic | 1318 | ±9 | 7,618 | proprietary |
| 🥈 | claude-opus-4-7-thinking | Anthropic | 1304 | ±7 | 19,578 | proprietary |
| 🥉 | gemini-3.6-flash | 1301 | ±38 | 238 | proprietary | |
| #4 | claude-opus-4-6-thinking | Anthropic | 1300 | ±7 | 19,445 | proprietary |
| #5 | claude-opus-5-high | Anthropic | 1299 | ±14 | 1,992 | proprietary |
| #6 | claude-opus-4-7 | Anthropic | 1298 | ±7 | 20,000 | proprietary |
| #7 | muse-spark | Meta | 1295 | ±9 | 5,584 | proprietary |
| #8 | claude-opus-4-6 | Anthropic | 1294 | ±7 | 23,722 | proprietary |
| #9 | gemini-3-pro | 1289 | ±8 | 13,089 | proprietary | |
| #10 | gpt-5.5 | OpenAI | 1287 | ±7 | 17,752 | proprietary |
| #11 | claude-opus-4-8-thinking | Anthropic | 1286 | ±8 | 10,268 | proprietary |
| #12 | gemini-3.5-flash-medium | 1285 | ±10 | 4,657 | proprietary | |
| #13 | gpt-5.5-high | OpenAI | 1285 | ±7 | 16,623 | proprietary |
| #14 | muse-spark-1.1 | Meta | 1283 | ±11 | 3,776 | proprietary |
| #15 | gpt-5.4-high | OpenAI | 1283 | ±7 | 19,622 | proprietary |
| #16 | grok-4.5 | SpaceXAI | 1282 | ±12 | 2,839 | proprietary |
| #17 | claude-opus-4-8 | Anthropic | 1279 | ±8 | 10,626 | proprietary |
| #18 | gpt-5.4 | OpenAI | 1278 | ±7 | 19,583 | proprietary |
| #19 | gpt-5.2-chat-latest-20260210 | OpenAI | 1278 | ±7 | 15,403 | proprietary |
| #20 | gemini-3.1-pro-preview | 1278 | ±6 | 35,988 | proprietary |
🔍 Search 32 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | claude-opus-4-6-search | Anthropic | 1253 | ±5 | 112,201 | proprietary |
| 🥈 | gpt-5.5-search | OpenAI | 1240 | ±5 | 67,573 | proprietary |
| 🥉 | claude-fable-5 | Anthropic | 1237 | ±8 | 20,630 | proprietary |
| #4 | claude-opus-4-7 | Anthropic | 1233 | ±5 | 68,515 | proprietary |
| #5 | ernie-5.1 | Baidu | 1226 | ±10 | 3,815 | proprietary |
| #6 | claude-sonnet-4-6-search | Anthropic | 1221 | ±5 | 111,959 | proprietary |
| #7 | gemini-3.1-pro-grounding | 1212 | ±5 | 89,973 | proprietary | |
| #8 | gemini-3-pro-grounding | 1207 | ±5 | 37,255 | proprietary | |
| #9 | gpt-5.2-search | OpenAI | 1206 | ±6 | 52,718 | proprietary |
| #10 | grok-4.20-multi-agent-beta-0309 | SpaceXAI | 1205 | ±5 | 87,118 | proprietary |
| #11 | claude-opus-4-8 | Anthropic | 1205 | ±6 | 48,610 | proprietary |
| #12 | gpt-5.1-search | OpenAI | 1199 | ±5 | 60,019 | proprietary |
| #13 | gemini-3-flash-grounding | 1197 | ±5 | 125,928 | proprietary | |
| #14 | gpt-5.4-search | OpenAI | 1196 | ±5 | 87,364 | proprietary |
| #15 | grok-4.20-beta1 | SpaceXAI | 1189 | ±6 | 53,877 | proprietary |
| #16 | claude-sonnet-5-search | Anthropic | 1188 | ±6 | 17,604 | proprietary |
| #17 | claude-opus-4-5-search | Anthropic | 1179 | ±6 | 61,611 | proprietary |
| #18 | gpt-5.2-search-non-reasoning | OpenAI | 1173 | ±5 | 75,752 | proprietary |
| #19 | grok-4-1-fast-search | SpaceXAI | 1171 | ±5 | 81,860 | proprietary |
| #20 | grok-4-fast-search | SpaceXAI | 1170 | ±4 | 42,981 | proprietary |
| #21 | grok-4.3 | SpaceXAI | 1163 | ±5 | 68,373 | proprietary |
| #22 | claude-sonnet-4-5-search | Anthropic | 1158 | ±5 | 105,762 | proprietary |
| #23 | claude-opus-4-1-search | Anthropic | 1148 | ±5 | 77,222 | proprietary |
| #24 | o3-search | OpenAI | 1143 | ±5 | 20,785 | proprietary |
| #25 | gemini-2.5-pro-grounding | 1142 | ±5 | 83,608 | proprietary | |
| #26 | grok-4-search | SpaceXAI | 1141 | ±6 | 19,379 | proprietary |
| #27 | ppl-sonar-reasoning-pro-high | Perplexity AI | 1138 | ±6 | 29,181 | proprietary |
| #28 | gpt-5-search | OpenAI | 1132 | ±6 | 20,913 | proprietary |
| #29 | ppl-sonar-pro-high | Perplexity AI | 1129 | ±6 | 28,703 | proprietary |
| #30 | claude-opus-4-search | Anthropic | 1126 | ±5 | 31,205 | proprietary |
| #31 | diffbot-small-xl | Diffbot | 1023 | ±8 | 6,433 | open |
| #32 | api-gpt-4o-search | OpenAI | 1006 | ±11 | 3,437 | proprietary |
📄 Document 35 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | claude-opus-5-high | Anthropic | 1520 | ±15 | 1,663 | proprietary |
| 🥈 | claude-opus-4-6 | Anthropic | 1510 | ±6 | 37,271 | proprietary |
| 🥉 | claude-opus-4-6-thinking | Anthropic | 1506 | ±7 | 24,973 | proprietary |
| #4 | claude-fable-5 | Anthropic | 1504 | ±9 | 5,792 | proprietary |
| #5 | claude-opus-4-7 | Anthropic | 1498 | ±7 | 18,990 | proprietary |
| #6 | claude-opus-4-7-thinking | Anthropic | 1497 | ±7 | 18,793 | proprietary |
| #7 | gpt-5.5-high | OpenAI | 1485 | ±7 | 16,816 | proprietary |
| #8 | claude-sonnet-4-6 | Anthropic | 1483 | ±6 | 54,427 | proprietary |
| #9 | gpt-5.5 | OpenAI | 1480 | ±7 | 17,286 | proprietary |
| #10 | claude-opus-4-8-thinking | Anthropic | 1475 | ±8 | 8,388 | proprietary |
| #11 | muse-spark-1.1 | Meta | 1472 | ±13 | 2,054 | proprietary |
| #12 | claude-sonnet-5-high | Anthropic | 1470 | ±10 | 3,601 | proprietary |
| #13 | gpt-5.4 | OpenAI | 1470 | ±7 | 29,809 | proprietary |
| #14 | claude-opus-4-8 | Anthropic | 1469 | ±8 | 8,193 | proprietary |
| #15 | gemini-3.5-flash-medium | 1465 | ±10 | 3,689 | proprietary | |
| #16 | claude-opus-4-5-20251101 | Anthropic | 1462 | ±10 | 7,965 | proprietary |
| #17 | grok-4.5 | SpaceXAI | 1454 | ±12 | 2,435 | proprietary |
| #18 | kimi-k2.6 | Moonshot | 1451 | ±8 | 11,181 | open |
| #19 | claude-sonnet-4-5-20250929 | Anthropic | 1446 | ±7 | 28,634 | proprietary |
| #20 | gemini-3.1-pro-preview | 1445 | ±6 | 45,162 | proprietary | |
| #21 | muse-spark | Meta | 1443 | ±18 | 1,078 | proprietary |
| #22 | qwen3.7-plus | Alibaba | 1440 | ±11 | 3,088 | proprietary |
| #23 | minimax-m3 | MiniMax | 1435 | ±8 | 6,263 | open |
| #24 | gemini-3-pro | 1433 | ±9 | 10,739 | proprietary | |
| #25 | kimi-k2.5-thinking | Moonshot | 1429 | ±7 | 19,891 | open |
| #26 | gemma-4-31b | 1424 | ±8 | 10,577 | open | |
| #27 | claude-haiku-4-5-20251001 | Anthropic | 1423 | ±6 | 30,937 | proprietary |
| #28 | gemini-2.5-pro | 1421 | ±6 | 24,963 | proprietary | |
| #29 | glm-5v-turbo | Z.ai | 1417 | ±10 | 4,894 | proprietary |
| #30 | grok-4.20-beta-0309-reasoning | SpaceXAI | 1414 | ±7 | 18,666 | proprietary |
| #31 | gemini-3-flash | 1413 | ±9 | 7,173 | proprietary | |
| #32 | gpt-5.2-high | OpenAI | 1405 | ±10 | 7,073 | proprietary |
| #33 | gpt-5.5-instant | OpenAI | 1402 | ±8 | 8,442 | proprietary |
| #34 | gpt-5.1 | OpenAI | 1401 | ±9 | 8,220 | proprietary |
| #35 | gpt-5.2 | OpenAI | 1400 | ±6 | 28,068 | proprietary |
🎨 Text → Image 74 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | gpt-image-2 (medium) | OpenAI | 1385 | ±5 | 60,382 | proprietary |
| 🥈 | reve-2.1 | Reve | 1302 | ±12 | 2,432 | proprietary |
| 🥉 | muse-image | Meta | 1280 | ±8 | 8,384 | proprietary |
| #4 | reve-2.0 | Reve | 1271 | ±6 | 13,675 | proprietary |
| #5 | gemini-3.1-flash-image (nano-banana-2) [web-search] | 1261 | ±7 | 18,502 | proprietary | |
| #6 | mai-image-2.5 | Microsoft AI | 1257 | ±5 | 34,416 | proprietary |
| #7 | gemini-3.1-flash-lite-image (nano-banana-2-lite) | 1250 | ±8 | 9,392 | proprietary | |
| #8 | gemini-3-pro-image-2k (nano-banana-pro) | 1245 | ±3 | 129,467 | proprietary | |
| #9 | gpt-image-1.5-high-fidelity | OpenAI | 1240 | ±3 | 133,306 | proprietary |
| #10 | gemini-3-pro-image-preview (nano-banana-pro) | 1232 | ±5 | 82,565 | proprietary | |
| #11 | seedream-5.0-pro | Bytedance | 1231 | ±11 | 2,865 | proprietary |
| #12 | grok-imagine-image-quality | SpaceXAI | 1229 | ±4 | 39,467 | proprietary |
| #13 | ideogram-4.0-quality | Ideogram | 1207 | ±6 | 18,547 | open |
| #14 | qwen-image-2.0-pro-2026-06-22 | Alibaba | 1193 | ±7 | 8,419 | proprietary |
| #15 | uni-1.1-max | Luma AI | 1188 | ±6 | 13,458 | proprietary |
| #16 | mai-image-2 | Microsoft AI | 1183 | ±5 | 49,043 | proprietary |
| #17 | uni-1.1 | Luma AI | 1177 | ±5 | 20,320 | proprietary |
| #18 | grok-imagine-image | SpaceXAI | 1173 | ±3 | 205,635 | proprietary |
| #19 | recraft-v4.1-utility-pro | Recraft | 1169 | ±11 | 2,510 | proprietary |
| #20 | flux-2-max | Black Forest Labs | 1162 | ±4 | 117,111 | proprietary |
| #21 | grok-imagine-image-pro | SpaceXAI | 1161 | ±4 | 93,439 | proprietary |
| #22 | flux-2-flex | Black Forest Labs | 1156 | ±3 | 148,925 | proprietary |
| #23 | flux-2-pro | Black Forest Labs | 1155 | ±3 | 166,182 | proprietary |
| #24 | reve-v1.5 | Reve | 1154 | ±4 | 31,694 | proprietary |
| #25 | gemini-2.5-flash-image-preview (nano-banana) | 1151 | ±3 | 807,662 | proprietary | |
| #26 | hunyuan-image-3.0 | Tencent | 1151 | ±3 | 172,769 | open |
| #27 | imagen-ultra-4.0-generate-001 | 1148 | ±4 | 388,009 | proprietary | |
| #28 | flux-2-dev | Black Forest Labs | 1148 | ±5 | 62,500 | open |
| #29 | seedream-4.5 | Bytedance | 1147 | ±3 | 220,462 | proprietary |
| #30 | seedream-4-2k | Bytedance | 1141 | ±7 | 12,605 | proprietary |
| #31 | wan2.6-t2i | Alibaba | 1134 | ±3 | 171,024 | proprietary |
| #32 | seedream-5.0-lite | Bytedance | 1132 | ±4 | 87,149 | proprietary |
| #33 | recraft-v4.1-pro | Recraft | 1130 | ±11 | 2,674 | proprietary |
| #34 | imagen-4.0-generate-001 | 1129 | ±3 | 531,357 | proprietary | |
| #35 | qwen-image-2512 | Alibaba | 1128 | ±4 | 83,631 | open |
| #36 | krea-2-medium | Krea | 1120 | ±6 | 17,491 | proprietary |
| #37 | hidream-o1-image | HiDream | 1118 | ±5 | 24,852 | open |
| #38 | seedream-4-fal | Bytedance | 1117 | ±7 | 11,851 | proprietary |
| #39 | wan2.5-t2i-preview | Alibaba | 1117 | ±3 | 219,089 | proprietary |
| #40 | gpt-image-1 | OpenAI | 1115 | ±3 | 264,724 | proprietary |
| #41 | krea-2-turbo | Krea | 1113 | ±7 | 9,143 | open |
| #42 | seedream-4-high-res-fal | Bytedance | 1113 | ±3 | 174,974 | proprietary |
| #43 | recraft-v4 | Recraft | 1113 | ±4 | 89,987 | proprietary |
| #44 | gpt-image-1-mini | OpenAI | 1109 | ±3 | 163,465 | proprietary |
| #45 | wan2.7-image-pro | Alibaba | 1102 | ±5 | 28,673 | proprietary |
| #46 | krea-2-large | Krea | 1100 | ±6 | 17,210 | proprietary |
| #47 | wan2.7-image | Alibaba | 1099 | ±5 | 29,016 | proprietary |
| #48 | mai-image-1 | Microsoft AI | 1093 | ±4 | 94,383 | proprietary |
| #49 | seedream-3 | Bytedance | 1082 | ±5 | 36,846 | proprietary |
| #50 | z-image-turbo | Alibaba | 1082 | ±6 | 19,864 | open |
| #51 | flux-1-kontext-max | Black Forest Labs | 1074 | ±3 | 65,669 | proprietary |
| #52 | flux-2-klein-9b | Black Forest Labs | 1069 | ±3 | 142,277 | open |
| #53 | qwen-image-prompt-extend | Alibaba | 1060 | ±3 | 704,537 | open |
| #54 | flux-1-kontext-pro | Black Forest Labs | 1059 | ±3 | 330,789 | proprietary |
| #55 | imagen-3.0-generate-002 | 1058 | ±3 | 360,193 | proprietary | |
| #56 | Cosmos3-Super-Text2Image | Nvidia | 1057 | ±6 | 18,247 | open |
| #57 | Cosmos3-Super-Text2Image (Agentic) | Nvidia | 1057 | ±5 | 18,221 | open |
| #58 | qwen-image | Alibaba | 1057 | ±3 | 84,663 | open |
| #59 | ideogram-v3-quality | Ideogram | 1049 | ±4 | 115,137 | proprietary |
| #60 | photon | Luma AI | 1035 | ±4 | 127,404 | proprietary |
| #61 | p-image | 1034 | ±4 | 103,638 | proprietary | |
| #62 | flux-2-klein-4b | Black Forest Labs | 1030 | ±3 | 144,073 | open |
| #63 | runway-gen4 | Runway | 1025 | ±5 | 51,604 | proprietary |
| #64 | recraft-v3 | Recraft | 1021 | ±4 | 191,873 | proprietary |
| #65 | flux-1.1-pro | Black Forest Labs | 1016 | ±3 | 70,460 | proprietary |
| #66 | lucid-origin | Leonardo AI | 1013 | ±3 | 285,498 | proprietary |
| #67 | ideogram-v2 | Ideogram | 1013 | ±4 | 72,090 | proprietary |
| #68 | glm-image | Z.ai | 1011 | ±9 | 4,612 | open |
| #69 | gemini-2.0-flash-preview-image-generation | 975 | ±3 | 257,095 | proprietary | |
| #70 | flux-1-dev-fp8 | Black Forest Labs | 970 | ±4 | 49,221 | open |
| #71 | dall-e-3 | OpenAI | 968 | ±4 | 239,288 | proprietary |
| #72 | flux-1-kontext-dev | Black Forest Labs | 940 | ±4 | 215,355 | open |
| #73 | stable-diffusion-v35-large | 938 | ±4 | 23,396 | open | |
| #74 | bagel | Bytedance | 898 | ±6 | 12,423 | open |
🎬 Text → Video 42 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | gemini-omni-flash | 1527 | ±13 | 5,449 | proprietary | |
| 🥈 | dreamina-seedance-2.0-720p | Bytedance | 1482 | ±10 | 41,953 | proprietary |
| 🥉 | muse-video | Meta | 1459 | ±15 | 2,152 | proprietary |
| #4 | happyhorse-1.0 | Alibaba-ATH | 1430 | ±13 | 21,985 | proprietary |
| #5 | sora-2-pro | OpenAI | 1366 | ±8 | 39,773 | proprietary |
| #6 | veo-3.1-audio-1080p | 1364 | ±11 | 23,200 | proprietary | |
| #7 | veo-3.1-audio | 1364 | ±14 | 13,691 | proprietary | |
| #8 | veo-3.1-fast-audio | 1362 | ±11 | 39,333 | proprietary | |
| #9 | veo-3.1-fast-audio-1080p | 1360 | ±10 | 23,506 | proprietary | |
| #10 | grok-imagine-video-720p | SpaceXAI | 1352 | ±8 | 144,437 | proprietary |
| #11 | wan2.7-t2v | Alibaba | 1348 | ±10 | 8,092 | proprietary |
| #12 | veo-3-fast-audio | 1347 | ±11 | 25,153 | proprietary | |
| #13 | veo-3-audio | 1339 | ±12 | 18,961 | proprietary | |
| #14 | sora-2 | OpenAI | 1337 | ±7 | 53,767 | proprietary |
| #15 | wan2.6-t2v | Alibaba | 1333 | ±10 | 34,624 | proprietary |
| #16 | seedance-v1.5-pro | Bytedance | 1257 | ±7 | 72,436 | proprietary |
| #17 | wan2.5-t2v-preview | Alibaba | 1254 | ±11 | 22,102 | proprietary |
| #18 | veo-3 | 1252 | ±11 | 14,948 | proprietary | |
| #19 | veo-3-fast | 1248 | ±12 | 15,230 | proprietary | |
| #20 | pixverse-v5.6 | 1241 | ±11 | 30,216 | proprietary | |
| #21 | runway-gen-4.5 | Runway | 1228 | ±12 | 30,013 | proprietary |
| #22 | kling-2.5-turbo-1080p | KlingAI | 1219 | ±17 | 2,104 | proprietary |
| #23 | kling-2.6-pro | KlingAI | 1218 | ±7 | 70,615 | proprietary |
| #24 | p-video | 1207 | ±16 | 7,039 | proprietary | |
| #25 | kling-o1-pro | KlingAI | 1205 | ±27 | 1,193 | proprietary |
| #26 | ray-3 | Luma AI | 1205 | ±22 | 1,121 | proprietary |
| #27 | hailuo-2.3 | MiniMax | 1199 | ±7 | 62,088 | proprietary |
| #28 | hailuo-02-pro | MiniMax | 1197 | ±12 | 9,367 | proprietary |
| #29 | seedance-v1-pro | Bytedance | 1190 | ±11 | 12,120 | proprietary |
| #30 | hailuo-02-standard | MiniMax | 1179 | ±12 | 9,333 | proprietary |
| #31 | kandinsky-5.0-t2v-pro | Kandinsky | 1172 | ±20 | 2,018 | open |
| #32 | hunyuan-video-1.5 | Tencent | 1169 | ±16 | 4,271 | |
| #33 | veo-2 | 1163 | ±16 | 6,509 | proprietary | |
| #34 | kling-v2.1-master | KlingAI | 1162 | ±10 | 14,049 | proprietary |
| #35 | ltx-2-19b | 1143 | ±10 | 53,873 | ||
| #36 | wan-v2.2-a14b | Alibaba | 1131 | ±15 | 10,418 | open |
| #37 | kandinsky-5.0-t2v-lite | Kandinsky | 1113 | ±18 | 1,475 | open |
| #38 | seedance-v1-lite | Bytedance | 1112 | ±9 | 16,212 | proprietary |
| #39 | sora | OpenAI | 1068 | ±16 | 4,080 | proprietary |
| #40 | ray2 | Luma AI | 1064 | ±17 | 5,217 | proprietary |
| #41 | pika-v2.2 | Pika | 1008 | ±15 | 5,728 | proprietary |
| #42 | mochi-v1 | Genmo AI | 1005 | ±16 | 5,862 | open |
📽️ Image → Video 42 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | dreamina-seedance-2.0-720p | Bytedance | 1474 | ±10 | 81,746 | proprietary |
| 🥈 | gemini-omni-flash | 1469 | ±11 | 5,373 | proprietary | |
| 🥉 | grok-imagine-video-1.5-preview-720p | SpaceXAI | 1466 | ±5 | 44,400 | proprietary |
| #4 | happyhorse-1.0 | Alibaba-ATH | 1444 | ±10 | 61,051 | proprietary |
| #5 | wan2.7-i2v | Alibaba | 1434 | ±8 | 14,113 | proprietary |
| #6 | grok-imagine-video-720p | SpaceXAI | 1422 | ±5 | 462,748 | proprietary |
| #7 | veo-3.1-audio | 1398 | ±11 | 25,112 | proprietary | |
| #8 | veo-3.1-audio-1080p | 1391 | ±9 | 38,881 | proprietary | |
| #9 | veo-3.1-fast-audio | 1385 | ±9 | 99,804 | proprietary | |
| #10 | grok-imagine-video-480p | SpaceXAI | 1384 | ±9 | 19,417 | proprietary |
| #11 | veo-3.1-fast-audio-1080p | 1374 | ±10 | 39,253 | proprietary | |
| #12 | vidu-q3-pro | Shengshu | 1361 | ±8 | 36,677 | proprietary |
| #13 | kling-v3-pro | KlingAI | 1360 | ±8 | 115,181 | proprietary |
| #14 | veo-3-audio | 1330 | ±11 | 32,383 | proprietary | |
| #15 | veo-3-fast-audio | 1324 | ±9 | 41,215 | proprietary | |
| #16 | wan2.5-i2v-preview | Alibaba | 1323 | ±10 | 16,932 | proprietary |
| #17 | wan2.6-i2v | Alibaba | 1316 | ±10 | 76,951 | proprietary |
| #18 | seedance-v1.5-pro | Bytedance | 1308 | ±7 | 218,351 | proprietary |
| #19 | pixverse-v5.6 | 1300 | ±8 | 108,232 | proprietary | |
| #20 | kling-2.6-pro | KlingAI | 1294 | ±8 | 182,012 | proprietary |
| #21 | kling-2.5-turbo-1080p | KlingAI | 1275 | ±12 | 3,791 | proprietary |
| #22 | seedance-v1-pro | Bytedance | 1272 | ±7 | 34,027 | proprietary |
| #23 | hailuo-2.3 | MiniMax | 1260 | ±6 | 229,374 | proprietary |
| #24 | veo-3-fast | 1257 | ±10 | 26,297 | proprietary | |
| #25 | veo-3 | 1256 | ±10 | 26,105 | proprietary | |
| #26 | p-video | 1244 | ±16 | 23,366 | proprietary | |
| #27 | vidu-q2-turbo | Shengshu | 1243 | ±17 | 2,506 | proprietary |
| #28 | kling-v2.1-master | KlingAI | 1234 | ±8 | 29,849 | proprietary |
| #29 | hailuo-02-pro | MiniMax | 1228 | ±10 | 21,751 | proprietary |
| #30 | kling-v2.1-standard | KlingAI | 1228 | ±8 | 29,952 | proprietary |
| #31 | ray-3 | Luma AI | 1225 | ±19 | 1,588 | proprietary |
| #32 | hailuo-02-standard | MiniMax | 1223 | ±9 | 21,782 | proprietary |
| #33 | vidu-q2-pro | Shengshu | 1222 | ±17 | 2,608 | proprietary |
| #34 | hunyuan-video-1.5 | Tencent | 1196 | ±15 | 5,476 | open |
| #35 | hailuo-02-fast | MiniMax | 1193 | ±10 | 22,549 | proprietary |
| #36 | seedance-v1-lite | Bytedance | 1184 | ±8 | 33,754 | proprietary |
| #37 | wan-v2.2-a14b | Alibaba | 1169 | ±10 | 27,067 | open |
| #38 | veo-2 | 1165 | ±16 | 10,319 | proprietary | |
| #39 | ltx-2-19b | 1147 | ±6 | 166,117 | open | |
| #40 | ray2 | Luma AI | 1107 | ±16 | 9,527 | proprietary |
| #41 | runway-gen4-turbo | Runway | 1051 | ±13 | 6,811 | proprietary |
| #42 | pika-v2.2 | Pika | 996 | ±13 | 8,655 | proprietary |
✏️ Image Edit 52 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | gpt-image-2 (medium) | OpenAI | 1465 | ±4 | 145,970 | proprietary |
| 🥈 | muse-image | Meta | 1402 | ±6 | 16,932 | proprietary |
| 🥉 | mai-image-2.5 | Microsoft AI | 1401 | ±4 | 100,526 | proprietary |
| #4 | seedream-5.0-pro | Bytedance | 1393 | ±10 | 4,299 | proprietary |
| #5 | chatgpt-image-latest-high-fidelity (20251216) | OpenAI | 1389 | ±3 | 457,021 | proprietary |
| #6 | grok-imagine-image-quality (20260519) | SpaceXAI | 1389 | ±5 | 34,756 | proprietary |
| #7 | gemini-3-pro-image-2k (nano-banana-pro) | 1388 | ±3 | 459,237 | proprietary | |
| #8 | gemini-3-pro-image-preview (nano-banana-pro) | 1385 | ±3 | 518,398 | proprietary | |
| #9 | gemini-3.1-flash-image (nano-banana-2) [web-search] | 1385 | ±4 | 60,941 | proprietary | |
| #10 | reve-2.1 | Reve | 1383 | ±9 | 5,165 | proprietary |
| #11 | gpt-image-1.5-high-fidelity | OpenAI | 1372 | ±3 | 478,581 | proprietary |
| #12 | grok-imagine-image-quality | SpaceXAI | 1359 | ±4 | 111,066 | proprietary |
| #13 | reve-2.0 | Reve | 1358 | ±7 | 14,513 | proprietary |
| #14 | uni-1.1-max | Luma AI | 1334 | ±5 | 40,859 | proprietary |
| #15 | grok-imagine-image | SpaceXAI | 1329 | ±3 | 478,004 | proprietary |
| #16 | qwen-image-2.0-pro-2026-06-22 | Alibaba | 1315 | ±5 | 32,739 | proprietary |
| #17 | gemini-3.1-flash-lite-image (nano-banana-2-lite) | 1314 | ±5 | 26,611 | proprietary | |
| #18 | uni-1.1 | Luma AI | 1311 | ±4 | 71,570 | proprietary |
| #19 | wan2.7-image-pro | Alibaba | 1302 | ±4 | 41,467 | proprietary |
| #20 | hunyuan-image-3.0-instruct | Tencent | 1302 | ±3 | 277,487 | |
| #21 | seedream-4.5 | Bytedance | 1301 | ±3 | 898,428 | proprietary |
| #22 | wan2.7-image | Alibaba | 1301 | ±4 | 42,318 | proprietary |
| #23 | gemini-2.5-flash-image-preview (nano-banana) | 1295 | ±2 | 11,005,375 | proprietary | |
| #24 | seedream-5.0-lite | Bytedance | 1294 | ±3 | 271,130 | proprietary |
| #25 | seedream-4-2k | Bytedance | 1271 | ±7 | 213,066 | proprietary |
| #26 | flux-2-max | Black Forest Labs | 1262 | ±3 | 352,694 | proprietary |
| #27 | reve-v1.1 | Reve | 1261 | ±3 | 631,952 | proprietary |
| #28 | kling-image-o1 | KlingAI | 1251 | ±4 | 135,159 | proprietary |
| #29 | flux-2-pro | Black Forest Labs | 1242 | ±3 | 436,098 | proprietary |
| #30 | qwen-image-edit | Alibaba | 1241 | ±3 | 1,983,543 | open |
| #31 | reve-v1 | Reve | 1234 | ±5 | 380,892 | proprietary |
| #32 | qwen-image-edit-2511 | Alibaba | 1233 | ±3 | 401,599 | open |
| #33 | wan2.6-image | Alibaba | 1230 | ±3 | 471,616 | proprietary |
| #34 | flux-2-flex | Black Forest Labs | 1225 | ±3 | 415,399 | proprietary |
| #35 | flux-2-klein-9b | Black Forest Labs | 1224 | ±3 | 506,759 | proprietary |
| #36 | flux-2-dev | Black Forest Labs | 1220 | ±4 | 188,220 | proprietary |
| #37 | seedream-4-high-res-fal | Bytedance | 1217 | ±3 | 1,184,210 | proprietary |
| #38 | p-image-edit | 1211 | ±3 | 336,435 | proprietary | |
| #39 | seedream-4-fal | Bytedance | 1210 | ±6 | 153,950 | proprietary |
| #40 | reve-v1.1-fast | Reve | 1206 | ±3 | 521,652 | proprietary |
| #41 | reve-edit-fast | Reve | 1198 | ±4 | 221,315 | proprietary |
| #42 | flux-2-klein-4b | Black Forest Labs | 1188 | ±3 | 506,764 | open |
| #43 | flux-1-kontext-max | Black Forest Labs | 1181 | ±3 | 391,731 | proprietary |
| #44 | wan2.5-i2i-preview | Alibaba | 1181 | ±3 | 481,319 | proprietary |
| #45 | flux-1-kontext-pro | Black Forest Labs | 1176 | ±3 | 6,424,593 | proprietary |
| #46 | flux-1-kontext-dev | Black Forest Labs | 1149 | ±3 | 3,652,473 | proprietary |
| #47 | seededit-3.0 | Bytedance | 1139 | ±3 | 4,950,659 | proprietary |
| #48 | gpt-image-1 | OpenAI | 1139 | ±3 | 2,861,463 | proprietary |
| #49 | gpt-image-1-mini | OpenAI | 1124 | ±3 | 653,134 | proprietary |
| #50 | gemini-2.0-flash-preview-image-generation | 1081 | ±2 | 4,966,089 | proprietary | |
| #51 | bagel | Bytedance | 1026 | ±6 | 13,560 | open |
| #52 | step1x-edit | StepFun | 998 | ±4 | 155,733 | open |
🎞️ Video Edit 7 models
| Rank | Model | Vendor | ELO Score | ± CI | Votes | License |
|---|---|---|---|---|---|---|
| 🥇 | dreamina-seedance-2.0-720p | Bytedance | 1377 | ±12 | 3,010 | proprietary |
| 🥈 | gemini-omni-flash | 1347 | ±16 | 998 | proprietary | |
| 🥉 | happyhorse-1.0 | Alibaba-ATH | 1308 | ±12 | 2,681 | proprietary |
| #4 | grok-imagine-video | SpaceXAI | 1264 | ±8 | 11,736 | proprietary |
| #5 | kling-o3-pro | KlingAI | 1251 | ±10 | 6,165 | proprietary |
| #6 | kling-o1-pro | KlingAI | 1203 | ±9 | 8,322 | proprietary |
| #7 | runway-gen4-aleph | Runway | 1194 | ±8 | 8,368 | proprietary |