Best AI for writing: all 77 ranked models

Ordered by writing votes. Score shown: how often people preferred its answers to writing prompts. Higher is better, and neighbouring models are often too close to separate: check the range before reading much into a one-place gap.

Show the other 62 ranked models

#	Model	Position	Score	Rating	Source
16	GoGemini 3.6 FlashGoogle DeepMind	Ranked 16th of 77	1,46695% range 1,458 to 1,474	Top tier	Source for Gemini 3.6 Flash write something score (lmarena)
17	XiMiMo-V2.6-Pro Best open-weights option Open weightsXiaomi	Ranked 17th of 77	1,46495% range 1,443 to 1,485	Top tier	Source for MiMo-V2.6-Pro write something score (lmarena)
18	MeMuse Spark 1.3Meta	Ranked 18th of 77	1,46495% range 1,454 to 1,473	Top tier	Source for Muse Spark 1.3 write something score (lmarena)
19	AlQwen3.5-Max-PreviewAlibaba	Ranked 19th of 77	1,46495% range 1,453 to 1,474	Top tier	Source for Qwen3.5-Max-Preview write something score (lmarena)
20	ZaGLM-5.2 Open weightsZ.ai (Zhipu AI)	Ranked 20th of 77	1,46195% range 1,454 to 1,468	Top tier	Source for GLM-5.2 write something score (lmarena)
21	ZaGLM-5.3 Open weightsZ.ai (Zhipu AI)	Ranked 21st of 77	1,45995% range 1,451 to 1,468	Strong	Source for GLM-5.3 write something score (lmarena)
22	MeMuse SparkMeta	Ranked 22nd of 77	1,45995% range 1,445 to 1,472	Strong	Source for Muse Spark write something score (lmarena)
23	MoKimi K3 Open weightsMoonshot	Ranked 23rd of 77	1,45795% range 1,449 to 1,465	Strong	Source for Kimi K3 write something score (lmarena)
24	GoGemini 3 FlashGoogle DeepMind	Ranked 24th of 77	1,45795% range 1,447 to 1,466	Strong	Source for Gemini 3 Flash write something score (lmarena)
25	OpGPT-5.5OpenAI	Ranked 25th of 77	1,45595% range 1,449 to 1,462	Strong	Source for GPT-5.5 write something score (lmarena)
26	AnClaude Opus 4.8Anthropic	Ranked 26th of 77	1,45595% range 1,448 to 1,461	Strong	Source for Claude Opus 4.8 write something score (lmarena)
27	GoGemini 2.5 ProGoogle DeepMind	Ranked 27th of 77	1,45495% range 1,449 to 1,459	Strong	Source for Gemini 2.5 Pro write something score (lmarena)
28	ZaGLM-5.1 Open weightsZ.ai (Zhipu AI)	Ranked 28th of 77	1,45395% range 1,446 to 1,459	Strong	Source for GLM-5.1 write something score (lmarena)
29	OpGPT-5.6 SolOpenAI	Ranked 29th of 77	1,45095% range 1,443 to 1,458	Strong	Source for GPT-5.6 Sol write something score (lmarena)
30	AlQwen3.7-Max-PreviewAlibaba	Ranked 30th of 77	1,44695% range 1,420 to 1,473	Strong	Source for Qwen3.7-Max-Preview write something score (lmarena)
31	DeDeepSeek-V4-Pro Open weightsDeepSeek	Ranked 31st of 77	1,44695% range 1,439 to 1,454	Strong	Source for DeepSeek-V4-Pro write something score (lmarena)
32	AnClaude Opus 4.5 (20251101)Anthropic	Ranked 32nd of 77	1,44395% range 1,436 to 1,449	Strong	Source for Claude Opus 4.5 (20251101) write something score (lmarena)
33	XaGrok 4.5xAI	Ranked 33rd of 77	1,44295% range 1,434 to 1,450	Strong	Source for Grok 4.5 write something score (lmarena)
34	ZaGLM-5.3-Flash Open weightsZ.ai (Zhipu AI)	Ranked 34th of 77	1,44195% range 1,434 to 1,449	Strong	Source for GLM-5.3-Flash write something score (lmarena)
35	DeDeepSeek-V4-Pro Preview Open weightsDeepSeek	Ranked 35th of 77	1,44195% range 1,434 to 1,448	Strong	Source for DeepSeek-V4-Pro Preview write something score (lmarena)
36	BaERNIE 5.1Baidu	Ranked 36th of 77	1,44195% range 1,432 to 1,449	Strong	Source for ERNIE 5.1 write something score (lmarena)
37	DeDeepSeek-V4-Pro (High, 20260813) Open weightsDeepSeek	Ranked 37th of 77	1,44095% range 1,426 to 1,453	Strong	Source for DeepSeek-V4-Pro (High, 20260813) write something score (lmarena)
38	OpGPT-5.4OpenAI	Ranked 38th of 77	1,43995% range 1,432 to 1,445	Strong	Source for GPT-5.4 write something score (lmarena)
39	XaGrok 4.20 Beta 1xAI	Ranked 39th of 77	1,43895% range 1,428 to 1,448	Strong	Source for Grok 4.20 Beta 1 write something score (lmarena)
40	AlQwen3.7-PlusAlibaba	Ranked 40th of 77	1,43895% range 1,430 to 1,445	Middle	Source for Qwen3.7-Plus write something score (lmarena)
41	ZaGLM-5 Open weightsZ.ai (Zhipu AI)	Ranked 41st of 77	1,43895% range 1,429 to 1,447	Middle	Source for GLM-5 write something score (lmarena)
42	XiMiMo-V2.5-Pro Open weightsXiaomi	Ranked 42nd of 77	1,43895% range 1,432 to 1,444	Middle	Source for MiMo-V2.5-Pro write something score (lmarena)
43	DeDeepSeek-V4.1-Flash-20260910 Open weightsDeepSeek	Ranked 43rd of 77	1,43795% range 1,425 to 1,449	Middle	Source for DeepSeek-V4.1-Flash-20260910 write something score (lmarena)
44	MeMuse Spark 1.2Meta	Ranked 44th of 77	1,43795% range 1,424 to 1,450	Middle	Source for Muse Spark 1.2 write something score (lmarena)
45	XaGrok 4.20 Multi-Agent Beta (0309)xAI	Ranked 45th of 77	1,43695% range 1,430 to 1,443	Middle	Source for Grok 4.20 Multi-Agent Beta (0309) write something score (lmarena)
46	MeMuse Spark 1.1Meta	Ranked 46th of 77	1,43595% range 1,428 to 1,443	Middle	Source for Muse Spark 1.1 write something score (lmarena)
47	AnClaude Sonnet 4.6Anthropic	Ranked 47th of 77	1,43595% range 1,429 to 1,442	Middle	Source for Claude Sonnet 4.6 write something score (lmarena)
48	AlQwen3.6-Max-PreviewAlibaba	Ranked 48th of 77	1,43595% range 1,412 to 1,457	Middle	Source for Qwen3.6-Max-Preview write something score (lmarena)
49	MoKimi K2.6 Open weightsMoonshot	Ranked 49th of 77	1,43395% range 1,425 to 1,441	Middle	Source for Kimi K2.6 write something score (lmarena)
50	XaGrok 4.20 Beta (Reasoning, 0309)xAI	Ranked 50th of 77	1,43295% range 1,425 to 1,439	Middle	Source for Grok 4.20 Beta (Reasoning, 0309) write something score (lmarena)
51	OpGPT-5.1OpenAI	Ranked 51st of 77	1,42795% range 1,419 to 1,435	Middle	Source for GPT-5.1 write something score (lmarena)
52	XaGrok 4.6xAI	Ranked 52nd of 77	1,42695% range 1,418 to 1,434	Middle	Source for Grok 4.6 write something score (lmarena)
53	BaERNIE 5.0 (0110)Baidu	Ranked 53rd of 77	1,42695% range 1,418 to 1,434	Middle	Source for ERNIE 5.0 (0110) write something score (lmarena)
54	OpGPT-6.1 SolOpenAI	Ranked 54th of 77	1,42595% range 1,411 to 1,439	Middle	Source for GPT-6.1 Sol write something score (lmarena)
55	MoKimi K2.5 Open weightsMoonshot	Ranked 55th of 77	1,42295% range 1,416 to 1,428	Middle	Source for Kimi K2.5 write something score (lmarena)
56	GoGemini 3.5 Flash-LiteGoogle DeepMind	Ranked 56th of 77	1,42195% range 1,413 to 1,429	Middle	Source for Gemini 3.5 Flash-Lite write something score (lmarena)
57	OpGPT-6 AstraOpenAI	Ranked 57th of 77	1,42095% range 1,409 to 1,431	Middle	Source for GPT-6 Astra write something score (lmarena)
58	OpGPT-5.5 InstantOpenAI	Ranked 58th of 77	1,41895% range 1,408 to 1,427	Middle	Source for GPT-5.5 Instant write something score (lmarena)
59	GoGemma 4 31B Open weightsGoogle DeepMind	Ranked 59th of 77	1,41695% range 1,398 to 1,435	Behind the leaders	Source for Gemma 4 31B write something score (lmarena)
60	AnClaude Sonnet 5Anthropic	Ranked 60th of 77	1,41695% range 1,409 to 1,423	Behind the leaders	Source for Claude Sonnet 5 write something score (lmarena)
61	OpGPT-5.6 TerraOpenAI	Ranked 61st of 77	1,41395% range 1,406 to 1,420	Behind the leaders	Source for GPT-5.6 Terra write something score (lmarena)
62	ByDola Seed 2.0 ProByteDance	Ranked 62nd of 77	1,40695% range 1,400 to 1,412	Behind the leaders	Source for Dola Seed 2.0 Pro write something score (lmarena)
63	XiMiMo-V2.6-Flash Open weightsXiaomi	Ranked 63rd of 77	1,40595% range 1,393 to 1,416	Behind the leaders	Source for MiMo-V2.6-Flash write something score (lmarena)
64	AlQwen3.6-PlusAlibaba	Ranked 64th of 77	1,40395% range 1,395 to 1,411	Behind the leaders	Source for Qwen3.6-Plus write something score (lmarena)
65	AlQwen3.5-397B-A17B Open weightsAlibaba	Ranked 65th of 77	1,40295% range 1,397 to 1,408	Behind the leaders	Source for Qwen3.5-397B-A17B write something score (lmarena)
66	ZaGLM-4.7 Open weightsZ.ai (Zhipu AI)	Ranked 66th of 77	1,40195% range 1,388 to 1,414	Behind the leaders	Source for GLM-4.7 write something score (lmarena)
67	NvNVIDIA-Nemotron-3-Ultra-550B-A55B (NVFP4) Open weightsNVIDIA	Ranked 67th of 77	1,39895% range 1,385 to 1,412	Behind the leaders	Source for NVIDIA-Nemotron-3-Ultra-550B-A55B (NVFP4) write something score (lmarena)
68	OpGPT-5.6 LunaOpenAI	Ranked 68th of 77	1,39695% range 1,389 to 1,403	Behind the leaders	Source for GPT-5.6 Luna write something score (lmarena)
69	XaGrok 4.7xAI	Ranked 69th of 77	1,38195% range 1,370 to 1,392	Behind the leaders	Source for Grok 4.7 write something score (lmarena)
70	XaGrok 4.3xAI	Ranked 70th of 77	1,38095% range 1,373 to 1,386	Behind the leaders	Source for Grok 4.3 write something score (lmarena)
71	OpGPT-6 SolOpenAI	Ranked 71st of 77	1,37895% range 1,366 to 1,389	Behind the leaders	Source for GPT-6 Sol write something score (lmarena)
72	Amamazon-nova-experimental-chat (26-02-10)Amazon	Ranked 72nd of 77	1,36995% range 1,343 to 1,394	Behind the leaders	Source for amazon-nova-experimental-chat (26-02-10) write something score (lmarena)
73	OpGPT-6 LunaOpenAI	Ranked 73rd of 77	1,36295% range 1,356 to 1,368	Behind the leaders	Source for GPT-6 Luna write something score (lmarena)
74	MiMistral Large 4Mistral AI	Ranked 74th of 77	1,36195% range 1,349 to 1,373	Behind the leaders	Source for Mistral Large 4 write something score (lmarena)
75	AlQwen3.5-FlashAlibaba	Ranked 75th of 77	1,34395% range 1,336 to 1,350	Behind the leaders	Source for Qwen3.5-Flash write something score (lmarena)
76	AnClaude Haiku 5.5Anthropic	Ranked 76th of 77	1,33695% range 1,321 to 1,351	Behind the leaders	Source for Claude Haiku 5.5 write something score (lmarena)
77	ThInkling-Small Open weightsThinking Machines	Ranked 77th of 77	1,33695% range 1,328 to 1,344	Behind the leaders	Source for Inkling-Small write something score (lmarena)

Test: Arena text: creative writing. Source: lmarena. Data last changed 29 Oct 2026. How we test.