模型排行

排名	变化	模型	得分	置信区间
1	—	Gemini Omni Flash闭源Google	1,325	±11
2	—	HappyHorse-1.0闭源Alibaba-ATH	1,285	±9
3	—	HappyHorse-1.1闭源Alibaba-ATH	1,272	±10
4	—	Dreamina Seedance 2.0 720p闭源ByteDance Seed	1,271	±8
5	▲ 1	Wan2.7-260612闭源Alibaba	1,243	±10
6	▼ 1	Kling 3.0 1080p (Pro)闭源KlingAI	1,242	±8
7	—	Kling 3.0 Omni 1080p (Pro)闭源KlingAI	1,229	±8
8	—	grok-imagine-video闭源SpaceXAI	1,226	±8
9	—	Bach-1.0 Preview闭源Video Rebirth	1,220	±9
10	▲ 3	Wan 2.7闭源Alibaba	1,219	±10
11	▼ 1	Vidu Q3 Pro闭源Vidu	1,219	±8
12	▼ 1	Veo 3.1 Preview闭源Google	1,217	±8
13	▲ 2	Runway Gen-4.5闭源Runway	1,217	±8
14	▼ 2	PixVerse V5.6闭源PixVerse	1,217	±8
15	▼ 1	Kling 3.0 Omni 720p (Standard)闭源KlingAI	1,217	±8
16	—	Veo 3 Preview闭源Google	1,215	±8
17	▲ 1	Veo 3闭源Google	1,215	±8
18	▲ 1	SkyReels V4闭源Skywork AI	1,215	±10
19	▼ 2	Veo 3.1 Fast Preview闭源Google	1,214	±8
20	—	PixVerse V6闭源PixVerse	1,213	±8
21	—	Kling 3.0 720p (Standard)闭源KlingAI	1,211	±8
22	—	Veo 3.1 Lite闭源Google	1,207	±9
23	▲ 1	Veo 3.1闭源Google	1,205	±8
24	▼ 1	Kling 2.5 Turbo 1080p闭源KlingAI	1,204	±8
25	—	Veo 3.1 Fast闭源Google	1,202	±8
26	—	Kling O1 Pro (January)闭源KlingAI	1,200	±8
27	—	Kling 2.6 Pro (January)闭源KlingAI	1,199	±8
28	—	PixVerse V5.5闭源PixVerse	1,198	±9
29	▲ 1	Kling O1 Pro (December)闭源KlingAI	1,193	±8
30	▼ 1	Kling O1 Standard (December)闭源KlingAI	1,191	±8
31	▲ 2	Kling 2.6 Pro (December)闭源KlingAI	1,190	±8
32	▼ 1	Kling O1 Standard (January)闭源KlingAI	1,189	±8
33	▲ 2	Wan 2.6闭源Alibaba	1,188	±8
34	▼ 2	Ray 3闭源Luma Labs	1,187	±8
35	▼ 1	Kling 2.6 Standard (January)闭源KlingAI	1,187	±8
36	—	Sora 2 Pro闭源OpenAI	1,183	±8
37	—	Veo 3 Fast Preview闭源Google	1,179	±8
38	▲ 1	Hailuo 2.3闭源MiniMax	1,173	±8
39	▲ 1	Hailuo 02 Standard闭源MiniMax	1,173	±9
40	▼ 2	Seedance 1.5 pro闭源ByteDance Seed	1,173	±8
41	▲ 1	PixVerse V5闭源PixVerse	1,170	±8
42	▼ 1	Sora 2 (December)闭源OpenAI	1,170	±9
43	—	Vidu Q2闭源Vidu	1,168	±8
44	—	Sora 2 (October)闭源OpenAI	1,164	±8
45	—	Wan 2.5 Preview闭源Alibaba	1,159	±8
46	—	Hailuo 02 Pro闭源MiniMax	1,152	±9
47	▲ 1	Seedance 1.0闭源ByteDance Seed	1,136	±8
48	▼ 1	Kling 2.1 Master闭源KlingAI	1,135	±9
49	—	PixVerse V5.6 (January)闭源PixVerse	1,133	±8
50	—	LTX-2 Pro开源Lightricks	1,127	±8
51	—	LTX-2.3 Fast开源Lightricks	1,125	±8
52	—	LTX-2 Fast开源Lightricks	1,121	±8
53	—	Veo 2闭源Google	1,120	±8
54	—	Wan 2.2 A14B开源Alibaba	1,110	±8
55	—	LTX-2.3 Pro开源Lightricks	1,107	±8
56	▲ 1	Seedance 1.0 Mini闭源ByteDance Seed	1,085	±8
57	▼ 1	Kling 2.0闭源KlingAI	1,083	±8
58	—	PixVerse V4.5闭源PixVerse	1,079	±9
59	—	Pika 2.5闭源Pika Art	1,078	±8
60	—	P-Video闭源Pruna AI	1,067	±8
61	—	Marey闭源Moonvalley	1,052	±9
62	—	Agnes-Video-V2.0闭源Sapiens AI	1,052	±10
63	—	Sora闭源OpenAI	1,043	±8
64	—	Kling 1.5 Pro闭源KlingAI	1,032	±8
65	—	Pika 2.0闭源Pika Art	1,026	±8
66	▲ 1	T2V-01闭源MiniMax	1,024	±8
67	▼ 1	T2V-01-Director闭源MiniMax	1,023	±8
68	—	HunyuanVideo-1.5 (Fal)开源Tencent	1,020	±9
69	▲ 1	Kling 1.6 Standard闭源KlingAI	1,018	±8
70	▼ 1	Kling 1.6 Pro闭源KlingAI	1,017	±8
71	—	Wan 2.1 14B开源Alibaba	1,014	±8
72	—	Motion 2.0闭源Leonardo.Ai	1,013	±10
73	—	Vidu Q1闭源Vidu	1,006	±10
74	—	Mochi 1开源Genmo	1,000	±0
75	—	Hunyuan Video (Fal)开源Tencent	998	±8
76	—	Runway Gen 3 Alpha闭源Runway	984	±8
77	—	Ray 1闭源Luma Labs	974	±8
78	—	Krea Realtime开源Krea	966	±9
79	—	Wan 2.2 5B开源Alibaba	951	±9
80	—	Kling 1.0闭源KlingAI	951	±8
81	—	Pika 2.2闭源Pika Art	948	±9
82	—	Ray 2闭源Luma Labs	947	±9
83	▲ 1	Haiper 2.0闭源Haiper	939	±8
84	▼ 1	Pika 1.5闭源Pika Art	937	±8
85	—	Step-Video-T2V闭源StepFun	917	±10
86	—	CogVideoX-5B开源Z AI	807	±10
87	—	Pyramid Flow开源Open Source	765	±10

其余 37 个模型

智力进步

每家厂商当时最强的模型在 Artificial Analysis 智力指数上的分数，按模型发布的月份画。

涨得最多 · Anthropic +57.2目前最高 · Anthropic 64.6最高与最低相差 18.1 分时间范围 2025年11月 – 2026年11月

64.6Anthropic60.0Google56.9OpenAI50.1Meta46.5Kimi