模型排行

排名	变化	模型	得分	置信区间
1	—	GPT Image 2 (high)闭源OpenAI	1,340	±7
2	新	Reve 2.1闭源Reve	1,299	±11
3	▼ 1	MAI-Image-2.5闭源Microsoft AI	1,270	±9
4	▲ 1	GPT Image 1.5 (high)闭源OpenAI	1,263	±7
4	▲ 2	Nano Banana 2 (Gemini 3.1 Flash Image Preview)闭源Google	1,263	±7
6	▼ 2	Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)闭源Google	1,263	±9
7	—	Riverflow 2.0闭源Sourceful	1,253	±8
8	▼ 5	HiDream-O1-Image-1.5闭源HiDream	1,246	±8
9	▼ 1	Reve 2.0闭源Reve	1,241	±10
10	新	Seedream 5.0 Pro闭源ByteDance	1,240	±9
11	▼ 2	Nano Banana Pro (Gemini 3 Pro Image)闭源Google	1,226	±7
12	▼ 1	MAI-Image-2.5-Flash闭源Microsoft AI	1,208	±9
13	—	Recraft V4.1 Utility闭源Recraft	1,205	±8
14	▼ 2	grok-imagine-image-quality闭源SpaceXAI	1,204	±7
15	▼ 5	Recraft V4.1 Utility Pro闭源Recraft	1,201	±7
16	▼ 1	FLUX.2 [max]闭源Black Forest Labs	1,197	±7
17	▼ 1	Seedream 4.0闭源ByteDance Seed	1,192	±7
18	▼ 1	Krea 2 Large闭源Krea	1,190	±7
19	▼ 5	Krea 2 Medium闭源Krea	1,190	±7
20	▼ 1	MAI-Image-2闭源Microsoft AI	1,189	±7
21	新	Krea 2 Medium Turbo闭源Krea	1,186	±10
22	▼ 4	FLUX.2 [pro]闭源Black Forest Labs	1,186	±7
23	▼ 1	grok-imagine-image闭源SpaceXAI	1,182	±7
24	▼ 3	FLUX.2 [flex]闭源Black Forest Labs	1,181	±8
25	▼ 2	Qwen Image 2.0 Pro (2026-04-22)闭源Alibaba	1,174	±8
26	▼ 2	Luma UNI 1 Max闭源Luma Labs	1,174	±8
27	▼ 2	Imagen 4 Ultra闭源Google	1,172	±8
28	▼ 2	ImagineArt 2.0闭源ImagineArt	1,169	±7
29	▼ 2	Seedream 4.5闭源ByteDance Seed	1,169	±7
30	▼ 1	MAI-Image-2-Efficient闭源Microsoft AI	1,159	±7
31	—	Recraft V4.1 Pro闭源Recraft	1,157	±8
32	▼ 2	Nano Banana (Gemini 2.5 Flash Image)闭源Google	1,153	±7
33	▼ 1	Recraft V4.1闭源Recraft	1,153	±8
34	—	Luma UNI 1闭源Luma Labs	1,151	±8
35	▼ 2	ImagineArt 1.5 Preview闭源ImagineArt	1,148	±7
36	—	Wan 2.5 Preview闭源Alibaba	1,146	±8
37	—	Seedream 3.0闭源ByteDance Seed	1,145	±8
38	—	Wan2.6 Text to Image闭源Alibaba	1,140	±7
39	—	Recraft V4 Pro闭源Recraft	1,139	±8
40	▲ 3	GPT Image 1 (high)闭源OpenAI	1,138	±8
41	▼ 1	Wan 2.6 Image闭源Alibaba	1,137	±8
42	—	Recraft V4闭源Recraft	1,137	±8
43	▼ 2	Vivago 2.1闭源HiDream	1,135	±8
44	—	Kolors 2.1闭源KlingAI	1,129	±8
45	—	image-1闭源Api Airforce	1,125	±7
46	—	FLUX.1 Kontext [max]闭源Black Forest Labs	1,124	±8
47	▲ 2	Wan 2.7 Pro闭源Alibaba	1,120	±7
48	▼ 1	Lucid Origin Ultra闭源Leonardo.Ai	1,120	±8
49	▼ 1	Seedream 5.0 Lite闭源ByteDance	1,119	±8
50	▲ 1	Eigen Image闭源Eigen AI	1,110	±7
51	▲ 1	Wan 2.7闭源Alibaba	1,110	±7
52	▼ 2	Lucid Origin Fast闭源Leonardo.Ai	1,109	±8
53	—	Vivago 2.0闭源HiDream	1,107	±8
54	—	Imagen 3 (v002)闭源Google	1,107	±8
55	▲ 1	Qwen Image 2.0 (2026-03-03)闭源Alibaba	1,105	±8
56	▼ 1	Vidu Q2闭源Vidu	1,102	±7
57	—	Imagen 4 Standard闭源Google	1,101	±8
58	—	Reve Image (Halfmoon)闭源Reve	1,099	±8
59	—	FLUX1.1 [pro] Ultra闭源Black Forest Labs	1,098	±8
60	—	Kling Image 3.0 Omni闭源KlingAI	1,098	±7
61	▲ 1	GPT Image 1 Mini (medium)闭源OpenAI	1,093	±8
62	▼ 1	FLUX.1 Kontext [pro]闭源Black Forest Labs	1,093	±8
63	▲ 1	Dreamina 3.1闭源ByteDance	1,091	±8
64	▼ 1	FLUX1.1 [pro]闭源Black Forest Labs	1,089	±7
65	▲ 1	Qwen Image Plus 2601闭源Alibaba	1,080	±7
66	▲ 1	Ideogram 3.0闭源Ideogram	1,078	±8
67	▲ 1	Imagen 4 Fast闭源Google	1,075	±8
68	▲ 1	Ideogram v2闭源Ideogram	1,073	±7
69	▲ 1	P-Image闭源Pruna AI	1,072	±8
70	▲ 1	Midjourney v7 Alpha闭源Midjourney	1,069	±8
71	▲ 2	Midjourney v6闭源Midjourney	1,068	±7
72	—	FLUX.1 [pro]闭源Black Forest Labs	1,067	±7
73	▲ 1	Recraft V3闭源Recraft	1,067	±7
74	▲ 1	Luma Photon闭源Luma Labs	1,065	±8
75	▲ 1	Ideogram v2 Turbo闭源Ideogram	1,058	±7
76	▲ 1	Image-01闭源MiniMax	1,055	±8
77	▲ 1	Midjourney v6.1闭源Midjourney	1,053	±7
78	▲ 1	MAI Image 1闭源Microsoft AI	1,042	±8
79	▲ 1	Z-Image Base闭源Alibaba	1,037	±8
80	▲ 1	Ideogram v1闭源Ideogram	1,034	±8
81	▲ 1	Phoenix 1.0 Ultra闭源Leonardo.Ai	1,032	±8
82	▲ 2	Stable Diffusion 3 Large闭源Stability.ai	1,027	±7
83	▲ 2	Ideogram v2a闭源Ideogram	1,015	±8
84	▲ 2	Ideogram v2a Turbo闭源Ideogram	1,015	±8
85	▲ 3	Phoenix 0.9 Ultra闭源Leonardo.Ai	1,011	±8
86	▲ 1	Playground v3 (beta)闭源Playground AI	1,010	±8
87	▲ 3	Krea 1闭源Krea	1,005	±8
88	▲ 4	Phoenix 1.0 Fast闭源Leonardo.Ai	995	±8
89	▲ 4	Luma Photon Flash闭源Luma Labs	987	±8
90	▲ 4	Step Image Edit 2闭源StepFun	984	±10
91	▲ 5	Runway Gen-4 Image闭源Runway	977	±8
92	▲ 3	Recraft 20B闭源Recraft	976	±8
93	▲ 4	DALLE 3 HD闭源OpenAI	959	±8
94	▲ 5	DALLE 3闭源OpenAI	950	±8
95	▲ 6	Grok 2闭源SpaceXAI	924	±13
96	▲ 6	Amazon Titan G1 v2 (Standard)闭源Amazon	912	±8
97	▲ 6	Stable Diffusion 3 Large Turbo闭源Stability.ai	908	±8
98	▲ 6	Stable Diffusion 1.6闭源Stability.ai	903	±8
99	▲ 6	Amazon Titan G1 (Standard)闭源Amazon	900	±8
100	▲ 6	DALLE 2闭源OpenAI	744	±10

其余 50 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