LMArena WebDev

crowdrun by LMArenaunit: eloCC-BY-4.0observed Jul 21, 202650 modelssource ↗methodology ↗

Caveat: Head-to-head web app building votes.

#	Model	Score
1	Kimi K3Moonshot AI	1678 ±17· 1.8K votes
2	Claude Fable 5Anthropic	1634 ±12· 3K votes
3	Claude Opus 4.8thinkingAnthropic	1565 ±8· 8K votes
4	Claude Opus 4.7thinkingAnthropic	1559 ±7· 11.2K votes
5	Grok 4.5xAI	1557 ±12· 2.9K votes
6	Claude Sonnet 5high-effortAnthropic	1544 ±11· 3.6K votes
7	Claude Opus 4.6thinkingAnthropic	1542 ±6· 13.5K votes
8	Muse Spark 1.1Meta	1539 ±12· 2.7K votes
9	Gemini 3.6 FlashGoogle	1537 ±13· 2.3K votes
10	GLM-5.1Zhipu AI	1526 ±8· 6.5K votes
11	Claude Sonnet 4.6Anthropic	1523 ±6· 16.9K votes
12	Kimi K2.6Moonshot AI	1515 ±7· 9.4K votes
13	Qwen3.7 MaxAlibaba	1515 ±8· 7K votes
14	MiniMax-M3MiniMax (minimax.io)	1491 ±8· 7K votes
15	Claude Opus 4.5 (latest)thinking-32kAnthropic	1490 ±7· 13.1K votes
16	Qwen3.6 Max PreviewAlibaba	1480 ±12· 2.5K votes
17	Kimi K2.7 CodeMoonshot AI	1470 ±9· 4.7K votes
18	DeepSeek V4 ProthinkingDeepSeek	1459 ±7· 9.6K votes
19	Qwen3.6 PlusAlibaba	1458 ±6· 12.6K votes
20	Gemini 3.1 Pro PreviewGoogle	1444 ±5· 18.2K votes
21	GLM-4.7Zhipu AI	1440 ±10· 4.9K votes
22	Kimi K2.5thinkingMoonshot AI	1433 ±6· 15.7K votes
23	GLM-5Zhipu AI	1430 ±8· 7.5K votes
24	GPT-5.2OpenAI	1406 ±17· 1.5K votes
25	GLM-5V-TurboZhipu AI	1401 ±16· 1.6K votes
26	MiniMax-M2.7MiniMax (minimax.io)	1397 ±6· 11.2K votes
27	GPT-5.4 minihigh-effortOpenAI	1397 ±7· 10.7K votes
28	Qwen3.5 397B-A17BAlibaba	1396 ±6· 15.3K votes
29	Claude Sonnet 4.5 (latest)thinking-32kAnthropic	1388 ±7· 15.7K votes
30	GPT-5.4OpenAI	1386 ±19· 1K votes
31	Claude Opus 4.1 (latest)Anthropic	1386 ±9· 8.6K votes
32	MiniMax-M2.5MiniMax (minimax.io)	1381 ±8· 7.9K votes
33	DeepSeek V3.2thinkingDeepSeek	1368 ±8· 7.9K votes
34	Qwen3.5 122B-A10BAlibaba	1364 ±7· 8.2K votes
35	Grok 4.3xAI	1360 ±7· 9K votes
36	Qwen3.5 27BAlibaba	1357 ±8· 7.7K votes
37	GLM-4.6Zhipu AI	1356 ±9· 8.3K votes
38	GPT-5.1OpenAI	1339 ±7· 12.9K votes
39	GPT-5.2 CodexOpenAI	1334 ±8· 7.8K votes
40	GPT-5.1 CodexOpenAI	1330 ±10· 6.2K votes
41	Kimi K2 Thinking TurboMoonshot AI	1330 ±6· 15.4K votes
42	Claude Haiku 4.5 (latest)Anthropic	1327 ±5· 26.5K votes
43	MiniMax-M2MiniMax (minimax.io)	1305 ±9· 8.4K votes
44	Qwen3-Coder 480B-A35B InstructAlibaba	1281 ±7· 15.2K votes
45	Gemini 3.1 Flash Lite PreviewGoogle	1253 ±7· 13.6K votes
46	Qwen3.5 35B-A3BAlibaba	1250 ±16· 1.8K votes
47	GPT-5.1 Codex miniOpenAI	1240 ±18· 1.4K votes
48	Gemini 2.5 ProGoogle	1204 ±13· 3.3K votes
49	Mercury 2Inception	1164 ±23· 947 votes
50	Devstral MediumMistral	1093 ±23· 993 votes

Dot = elo on a 1038–1727 scale (position, not length); whisker = rating confidence interval. Scores link to model pages; every number keeps its source on the model page.