LMArena WebDev leaderboard
- models measured
- 101
- lead over #2
- +34 Elo
- median
- 1499
Leader on LMArena WebDev · read October 21
# LMArena WebDev, read October 21
leaders = {
"Claude Opus 5.5": 1807, "GPT-6 Astra": 1774, "Claude Sonnet 5.5": 1768,
}
On October 21, 2026, 101 of our catalog's models had a LMArena WebDev result, and Anthropic Claude Opus 5.5 led with 1807.
People compare two models building the same web page and vote for the better one. The votes become an Elo-style rating, so a higher number means the model wins more of those comparisons.
Role in the index
This benchmark carries 15% of the Vibe Coding Index weight. A model missing it is scored on the benchmarks it has. Ratings are rescaled from 1200 to 1900 onto 0 to 100. Read the methodology
| # | Model | Result | Votes | Interval (±) | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|---|
| 1 | Claude Opus 5.5Anthropic | 1807 | 2,925 | ±13.8 | 92.2 | $20.00 | |
| 2 | GPT-6 AstraOpenAI | 1774 | 6,169 | ±10.2 | 88.7 | $50.00 | |
| 3 | Claude Sonnet 5.5Anthropic | 1768 | 3,428 | ±12.8 | 88.9 | $10.00 | |
| 4 | GPT-6.1 SolOpenAI | 1753 | 3,072 | ±13.0 | 88.7 | $10.00 | |
| 5 | Claude Fable 5.1Anthropic | 1743 | 7,744 | ±8.9 | 89.1 | $50.00 | |
| 6 | Claude Opus 5Anthropic | 1689 | 18,555 | ±6.2 | 81.8 | $25.00 | |
| 7 | GPT-6 SolOpenAI | 1683 | 4,608 | ±10.4 | 82.8 | $10.00 | |
| 8 | Gemini 4 ArgonGoogle | 1678 | 1,728 | ±15.6 | 80.5 | $10.00 | |
| 9 | Qwen3.8 Max (0902)Qwen | 1674 | 14,421 | ±7.0 | 63.5 | $6.00 | |
| 10 | DeepSeek V4.1 Flash 1017DeepSeek | 1664 | 2,008 | ±14.3 | 83.3 | $0.60 | |
| 11 | Muse Spark 1.3Meta | 1659 | 9,407 | ±7.9 | 79.7 | $4.25 | |
| 12 | Kimi K3Moonshot AI | 1655 | 19,592 | ±6.0 | 61.8 | $11.936 | |
| 13 | Muse WonderMeta | 1652 | 2,289 | ±13.7 | 64.5 | $20.00 | |
| 14 | Grok 4.7xAI | 1637 | 5,303 | ±9.5 | 75.0 | $6.00 | |
| 15 | MiMo-V2.6-FlashXiaomi | 1634 | 5,293 | ±9.4 | 61.9 | $0.28 | |
| 16 | Hy4 previewTencent | 1631 | 4,798 | ±9.6 | 61.6 | $2.501 | |
| 17 | MiMo-V2.6-ProXiaomi | 1625 | 5,041 | ±9.6 | 72.6 | $0.87 | |
| 18 | Claude Fable 5Anthropic | 1624 | 15,899 | ±6.1 | 82.1 | $50.00 | |
| 19 | GLM 5.3Z.ai | 1621 | 9,828 | ±7.5 | 73.1 | $1.40 | |
| 20 | DeepSeek V4.1 FlashDeepSeek | 1618 | 3,528 | ±11.0 | 77.2 | $0.525 | |
| 21 | Grok 4.6xAI | 1617 | 10,483 | ±7.3 | 76.4 | $6.00 | |
| 22 | GPT-5.6 SolOpenAI | 1616 | 17,786 | ±5.9 | 76.7 | $20.00 | |
| 23 | GLM 5.3 FlashZ.ai | 1607 | 18,264 | ±6.6 | 71.0 | $0.50 | |
| 24 | GLM 5.2Z.ai | 1604 | 16,657 | ±6.0 | 57.7 | $7.00 | |
| 25 | Qwen3.8 27BQwen | 1593 | 16,385 | ±6.5 | 73.6 | $1.35 | |
| 26 | Gemini 3.7 FlashGoogle | 1591 | 12,636 | ±6.8 | 69.9 | $3.75 | |
| 27 | Claude Haiku 5.5Anthropic | 1586 | 3,514 | ±10.9 | 75.9 | $0.50 | |
| 28 | Gemini 3.8 FlashGoogle | 1585 | 10,851 | ±7.4 | 70.5 | $4.125 | |
| 29 | DeepSeek V4 Pro 0813DeepSeek | 1582 | 4,714 | ±9.5 | 54.6 | $1.98 | |
| 30 | DeepSeek V4 Flash 0423DeepSeek | 1582 | 4,947 | ±9.3 | 54.6 | $0.182 | |
| 31 | GPT-6 LunaOpenAI | 1578 | 21,622 | ±6.1 | 71.4 | $0.50 | |
| 32 | Step 5 PreviewStepFun | 1567 | 4,728 | ±9.8 | 71.4 | $2.70 | |
| 33 | Claude Opus 4.7Anthropic | 1558 | 17,910 | ±5.8 | 51.1 | $25.00 | |
| 34 | Claude Opus 4.8Anthropic | 1556 | 18,493 | ±5.8 | 66.0 | $25.00 | |
| 35 | Grok 4.5xAI | 1553 | 11,082 | ±6.7 | 75.5 | $6.00 | |
| 36 | Claude Opus 4.6Anthropic | 1547 | 19,383 | ±5.5 | 66.9 | $25.00 | |
| 37 | Mistral Large 4Mistral AI | 1543 | 2,832 | ±12.3 | 55.3 | $4.18 | |
| 38 | Muse Spark 1.1Meta | 1542 | 7,985 | ±7.6 | 48.9 | $4.25 | |
| 39 | Claude Sonnet 5Anthropic | 1540 | 15,755 | ±6.0 | 63.5 | $10.00 | |
| 40 | Gemini 3.6 FlashGoogle | 1538 | 8,762 | ±7.5 | 48.3 | $3.75 | |
| 41 | Muse Spark 1.2Meta | 1534 | 2,163 | ±13.7 | 65.1 | $4.25 | |
| 42 | Muse Spark OpenMeta | 1528 | 1,380 | ±17.0 | 62.1 | $4.25 | |
| 43 | Claude Sonnet 4.6Anthropic | 1523 | 22,822 | ±5.2 | 57.8 | $15.00 | |
| 44 | GPT-5.6 TerraOpenAI | 1522 | 13,919 | ±6.3 | 69.0 | $12.00 | |
| 45 | GPT-5.6 LunaOpenAI | 1518 | 12,501 | ±6.4 | 61.5 | $1.20 | |
| 46 | Qwen3.7 MaxQwen | 1515 | 8,887 | ±7.4 | 45.1 | $4.425 | |
| 47 | GPT-5.5OpenAI | 1513 | 15,612 | ±6.0 | 67.8 | $30.00 | |
| 48 | Kimi K2.6Moonshot AI | 1509 | 10,166 | ±7.2 | 44.2 | $3.41 | |
| 49 | GLM 5.1Z.ai | 1508 | 10,703 | ±6.8 | 44.0 | $4.40 | |
| 50 | Hy3Tencent | 1507 | 6,256 | ±8.4 | 43.9 | $0.528 | |
| 51 | Gemini 3.5 FlashGoogle | 1499 | 9,382 | ±7.3 | 42.8 | $9.00 | |
| 52 | Claude Opus 4.5Anthropic | 1494 | 11,029 | ±8.3 | 66.8 | $25.00 | |
| 53 | MiniMax M3MiniMax | 1482 | 19,803 | ±5.5 | 57.8 | $1.20 | |
| 54 | MiMo-V2.5-ProXiaomi | 1480 | 22,493 | ±5.2 | 40.0 | $0.609 | |
| 55 | Kimi K2.7 CodeMoonshot AI | 1474 | 4,962 | ±9.4 | 39.1 | $3.35 | |
| 56 | GPT-5.4OpenAI | 1464 | 1,350 | ±18.9 | 37.7 | $15.00 | |
| 57 | Qwen3.6 PlusQwen | 1461 | 18,587 | ±5.6 | 37.4 | $1.95 | |
| 58 | DeepSeek V4 Pro 0423DeepSeek | 1447 | 13,979 | ±6.3 | 35.3 | $1.9105 | |
| 59 | Gemini 3.1 Pro PreviewGoogle | 1447 | 24,199 | ±5.1 | 35.3 | $12.00 | |
| 60 | MiMo-V2.5Xiaomi | 1438 | 14,910 | ±6.0 | 34.0 | $0.238 | |
| 61 | R Beam 1.0 501B A23BReflection | 1437 | 2,930 | ±12.7 | 33.8 | $1.00 | |
| 62 | Kimi K2.5Moonshot AI | 1437 | 20,091 | ±5.2 | 58.5 | $2.50 | |
| 63 | GLM 5Z.ai | 1435 | 7,496 | ±8.2 | 59.7 | $3.20 | |
| 64 | GLM 4.7Z.ai | 1435 | 4,007 | ±12.0 | 33.6 | $1.98 | |
| 65 | Gemini 3.5 Flash LiteGoogle | 1435 | 238 | ±40.8 | 33.6 | $2.50 | |
| 66 | GPT-5OpenAI | 1418 | 3,092 | ±15.6 | 72.2 | $10.00 | |
| 67 | GPT-5.2OpenAI | 1417 | 1,081 | ±22.6 | 58.9 | $14.00 | |
| 68 | T InklingThinking Machines | 1414 | 15,778 | ±6.1 | 49.7 | $4.05 | |
| 69 | T Inkling SmallThinking Machines | 1409 | 5,323 | ±9.1 | 29.9 | $1.20 | |
| 70 | GPT-5.3-CodexOpenAI | 1407 | 2,641 | ±13.7 | 29.5 | $14.00 | |
| 71 | GLM 5V TurboZ.ai | 1402 | 2,403 | ±13.2 | 28.8 | $4.00 | |
| 72 | Qwen3.5 397B A17BQwen | 1400 | 21,156 | ±5.1 | 28.6 | $2.34 | |
| 73 | MiniMax M2.7MiniMax | 1399 | 15,972 | ±5.8 | 28.4 | $1.20 | |
| 74 | GPT-5.4 MiniOpenAI | 1397 | 11,942 | ±6.5 | 28.2 | $4.50 | |
| 75 | GPT-5.1OpenAI | 1396 | 4,910 | ±11.5 | 53.3 | $10.00 | |
| 76 | Claude Sonnet 4.5Anthropic | 1394 | 13,250 | ±7.7 | 59.1 | $15.00 | |
| 77 | Claude Opus 4.1Anthropic | 1390 | 6,927 | ±10.7 | 27.2 | $75.00 | |
| 78 | MiniMax M2.5MiniMax | 1386 | 7,621 | ±8.3 | 59.4 | $0.95 | |
| 79 | Solar Pro 4Upstage | 1371 | 4,655 | ±10.3 | 24.4 | $0.60 | |
| 80 | Gemma 4 31BGoogle | 1367 | 14,666 | ±6.2 | 38.3 | $0.40 | |
| 81 | DeepSeek V3.2DeepSeek | 1362 | 6,914 | ±9.1 | 54.4 | $0.42 | |
| 82 | Qwen3.5-122B-A10BQwen | 1361 | 8,235 | ±7.7 | 22.9 | $2.08 | |
| 83 | Gemma 4 26B A4BGoogle | 1360 | 1,501 | ±16.9 | 22.8 | $0.255 | |
| 84 | Qwen3.5-27BQwen | 1358 | 7,782 | ±8.0 | 22.6 | $2.16 | |
| 85 | Grok 4.3xAI | 1358 | 14,954 | ±6.2 | 22.5 | $2.50 | |
| 86 | Hy3 previewTencent | 1357 | 1,448 | ±17.4 | 22.4 | $0.60 | |
| 87 | GLM 4.6Z.ai | 1341 | 6,781 | ±10.8 | 52.2 | $2.20 | |
| 88 | GPT-5.2-CodexOpenAI | 1339 | 6,718 | ±8.9 | 55.1 | $14.00 | |
| 89 | GPT-5.1-CodexOpenAI | 1337 | 5,007 | ±11.6 | 50.5 | $10.00 | |
| 90 | Claude Haiku 4.5Anthropic | 1330 | 29,061 | ±4.9 | 49.8 | $5.00 | |
| 91 | MiniMax M2MiniMax | 1297 | 6,714 | ±10.9 | 45.3 | $1.02 | |
| 92 | Qwen3 Coder 480B A35BQwen | 1275 | 12,709 | ±7.8 | 50.0 | $1.00 | |
| 93 | DeepSeek V3.2 ExpDeepSeek | 1272 | 3,969 | ±13.5 | 55.0 | $0.41 | |
| 94 | Mistral Medium 3.5Mistral AI | 1264 | 2,317 | ±14.8 | 9.2 | $7.50 | |
| 95 | Gemini 3.1 Flash Lite PreviewGoogle | 1257 | 14,390 | ±6.8 | 8.1 | $1.50 | |
| 96 | Qwen3.5-35B-A3BQwen | 1254 | 1,617 | ±17.5 | 7.7 | $1.00 | |
| 97 | GPT-5.1-Codex-MiniOpenAI | 1244 | 1,154 | ±22.0 | 6.2 | $2.00 | |
| 98 | Qwen3.5-FlashQwen | 1244 | 1,328 | ±19.6 | 6.2 | $0.26 | |
| 99 | Trinity Large ThinkingArcee AI | 1238 | 1,361 | ±20.0 | 5.4 | $0.80 | |
| 100 | Gemini 2.5 ProGoogle | 1226 | 2,730 | ±15.9 | 37.0 | $10.00 | |
| 101 | Mercury 2Inception | 1171 | 954 | ±24.5 | 0.0 | $0.75 |
Only models in our catalog are listed, and each appears once with its best published run. 101 rows were read from the source on October 21, 2026, and rows that match no catalog model are left out. Every other column is a figure the board publishes for the same run.