LMArena Text

crowdrun by LMArenaunit: eloCC-BY-4.0observed Jul 21, 202659 modelssource ↗methodology ↗

Caveat: Overall text category; style effects not controlled in this config.

#	Model	Score
1	Claude Opus 4.6thinkingAnthropic	1512 ±4· 62.8K votes
2	Claude Fable 5Anthropic	1504 ±6· 14.6K votes
3	Claude Opus 4.7thinkingAnthropic	1499 ±4· 50.5K votes
4	GPT-5.4OpenAI	1499 ±5· 61.8K votes
5	GPT-5.4 minihigh-effortOpenAI	1496 ±5· 57.6K votes
6	GPT-5.2high-effortOpenAI	1493 ±7· 47.4K votes
7	Muse Spark 1.1Meta	1491 ±7· 7.8K votes
8	Gemini 3.6 FlashGoogle	1491 ±9· 4.7K votes
9	GPT-5.5OpenAI	1490 ±6· 47K votes
10	Gemini 3.5 Flashhigh-effortGoogle	1490 ±6· 10.1K votes
11	Gemini 3.1 Pro PreviewGoogle	1489 ±3· 84.1K votes
12	GPT-5.6 Solxhigh-effortOpenAI	1486 ±11· 6.1K votes
13	GPT-5.1OpenAI	1485 ±8· 42.5K votes
14	GPT-5.3 Chat (latest)OpenAI	1484 ±8· 32.8K votes
15	Kimi K3Moonshot AI	1482 ±10· 3.6K votes
16	GPT-5.4 nanohigh-effortOpenAI	1482 ±5· 56.6K votes
17	Grok 4.5xAI	1478 ±11· 7.7K votes
18	GLM-5.1Zhipu AI	1475 ±4· 30.5K votes
19	Claude Opus 4.8thinkingAnthropic	1472 ±5· 30.8K votes
20	Qwen3.7 PlusAlibaba	1468 ±5· 22.2K votes
21	Claude Sonnet 4.6Anthropic	1468 ±4· 56.8K votes
22	GPT-5.2 ChatOpenAI	1465 ±9· 34.1K votes
23	GPT-5high-effortOpenAI	1465 ±17· 6.8K votes
24	Gemini 2.5 FlashGoogle	1465 ±6· 68.7K votes
25	Kimi K2.6Moonshot AI	1465 ±4· 37.6K votes
26	GPT-5 Minihigh-effortOpenAI	1465 ±17· 6.8K votes
27	Claude Opus 4.5 (latest)thinking-32kAnthropic	1463 ±8· 36.2K votes
28	GPT-4.1OpenAI	1462 ±19· 7.8K votes
29	Gemini 2.5 ProGoogle	1462 ±3· 68K votes
30	DeepSeek V4 ProDeepSeek	1460 ±4· 45.1K votes
31	Claude Sonnet 5high-effortAnthropic	1460 ±11· 13.4K votes
32	Grok 4.3xAI	1457 ±6· 46.2K votes
33	Qwen3.6 Max PreviewAlibaba	1456 ±8· 5.2K votes
34	GLM-5Zhipu AI	1455 ±4· 27.6K votes
35	Kimi K2.5thinkingMoonshot AI	1455 ±3· 62.3K votes
36	Claude Opus 4.1 (latest)Anthropic	1454 ±8· 40.9K votes
37	GLM-5V-TurboZhipu AI	1450 ±7· 7.5K votes
38	Claude Sonnet 4.5 (latest)thinking-32kAnthropic	1450 ±6· 73K votes
39	GLM-4.6Zhipu AI	1449 ±9· 28.4K votes
40	Gemini 3.1 Flash Lite PreviewGoogle	1449 ±6· 60.5K votes
41	Qwen3.5 397B-A17BAlibaba	1448 ±4· 57.8K votes
42	Gemini 3.5 Flash LiteGoogle	1448 ±9· 4.6K votes
43	GLM-4.7Zhipu AI	1447 ±6· 11.7K votes
44	Qwen3.6 PlusAlibaba	1447 ±4· 43.8K votes
45	MiniMax-M3MiniMax (minimax.io)	1446 ±5· 28K votes
46	DeepSeek V3.2DeepSeek	1444 ±8· 46.3K votes
47	DeepSeek V4 FlashDeepSeek	1441 ±4· 44.9K votes
48	MiniMax-M2.5MiniMax (minimax.io)	1441 ±8· 40.8K votes
49	Command ACohere	1441 ±17· 7.6K votes
50	Claude Haiku 4.5 (latest)Anthropic	1439 ±5· 101.8K votes
51	MiniMax-M2.7MiniMax (minimax.io)	1435 ±6· 49.7K votes
52	GLM-4.6VZhipu AI	1433 ±29· 2.5K votes
53	Qwen3.5 122B-A10BAlibaba	1432 ±9· 28.3K votes
54	Qwen3.5 27BAlibaba	1429 ±10· 27.1K votes
55	GLM-4.5Zhipu AI	1429 ±5· 24.3K votes
56	DeepSeek R1DeepSeek	1428 ±6· 18.5K votes
57	Kimi K2 Thinking TurboMoonshot AI	1423 ±3· 61K votes
58	Qwen3-Next 80B-A3B InstructAlibaba	1423 ±7· 7.6K votes
59	DeepSeek V3.1DeepSeek	1419 ±6· 14.9K votes

Dot = elo on a 1398–1522 scale (position, not length); whisker = rating confidence interval. Scores link to model pages; every number keeps its source on the model page.