模型排行

排名	變化	模型	得分	置信區間	投票
1	新	gpt-image-2.5-sunburst閉源OpenAI	1,421	±13	3,149
2	新	gpt-image-2.5-flare閉源OpenAI	1,399	±13	2,856
3	▼ 2	gpt-image-2 (medium)閉源OpenAI	1,381	±4	78,731
4	▼ 2	mai-image-2.6閉源Microsoft AI	1,331	±7	11,213
5	▼ 2	grok-imagine-image-2.0 (low)閉源xAI	1,315	±12	2,681
6	▼ 2	reve-2.1閉源Reve	1,301	±8	7,573
7	▼ 2	muse-image閉源Meta	1,277	±6	26,045
8	▼ 2	reve-2.0閉源Reve	1,270	±6	14,632
9	▼ 2	gemini-3.1-flash-image (nano-banana-2) [web-search]閉源Google	1,261	±5	41,957
10	▼ 2	seedream-5.0-pro閉源ByteDance	1,257	±4	62,443
11	▼ 2	qwen-image-3.0-pro閉源Alibaba	1,254	±7	10,920
12	▼ 2	mai-image-2.5閉源Microsoft AI	1,254	±4	57,250
13	▼ 2	gemini-3.1-flash-lite-image (nano-banana-2-lite)閉源Google	1,250	±6	16,371
14	▼ 2	gemini-3-pro-image-2k (nano-banana-pro)閉源Google	1,246	±3	152,220
15	▼ 2	gpt-image-1.5-high-fidelity閉源OpenAI	1,239	±3	151,702
16	▼ 2	gemini-3-pro-image-preview (nano-banana-pro)閉源Google	1,232	±5	82,737
17	▼ 2	ideogram-4.0-quality開源Ideogram	1,204	±4	42,169
18	▼ 2	qwen-image-2.0-pro-2026-06-22閉源Alibaba	1,191	±6	12,005
19	▼ 2	uni-1.1-max閉源Luma AI	1,188	±6	13,469
20	▼ 2	mai-image-2閉源Microsoft AI	1,183	±5	49,198
21	▼ 2	uni-1.1閉源Luma AI	1,181	±5	33,836
22	▼ 2	grok-imagine-image閉源xAI	1,171	±3	236,623
23	▼ 2	recraft-v4.1-utility-pro閉源Recraft	1,169	±11	2,516
24	▼ 2	Cosmos3-Super-Text2Image (Agentic)開源NVIDIA	1,167	±8	7,424
25	▼ 2	flux-2-max閉源Black Forest Labs	1,162	±4	117,339
26	▼ 2	grok-imagine-image-pro閉源xAI	1,161	±4	93,639
27	▼ 2	flux-2-flex閉源Black Forest Labs	1,157	±3	149,288
28	▼ 2	flux-2-pro閉源Black Forest Labs	1,154	±3	188,083
29	▼ 2	reve-v1.5閉源Reve	1,154	±4	36,748
30	▼ 2	Cosmos3-Super-Text2Image開源NVIDIA	1,154	±7	12,499
31	▼ 2	hunyuan-image-3.0開源Tencent	1,151	±3	173,099
32	▼ 2	gemini-2.5-flash-image-preview (nano-banana)閉源Google	1,150	±2	845,086
33	▼ 2	imagen-ultra-4.0-generate-001閉源Google	1,148	±4	388,103
34	▼ 2	seedream-4.5閉源ByteDance	1,147	±3	259,145
35	▼ 2	flux-2-dev開源Black Forest Labs	1,145	±4	73,963
36	▼ 2	seedream-4-2k閉源ByteDance	1,140	±7	12,621
37	▼ 2	seedream-5.0-lite閉源ByteDance	1,138	±3	109,182
38	▼ 2	wan2.6-t2i閉源Alibaba	1,137	±3	201,108
39	▼ 2	recraft-v4.1-pro閉源Recraft	1,130	±11	2,680
40	▼ 2	imagen-4.0-generate-001閉源Google	1,129	±3	531,811
41	▼ 2	qwen-image-2512開源Alibaba	1,125	±4	92,541
42	▼ 2	krea-2-medium閉源Krea	1,123	±5	40,205
43	▼ 2	wan2.5-t2i-preview閉源Alibaba	1,118	±3	248,411
44	▼ 2	hidream-o1-image開源HiDream	1,117	±4	48,201
45	▼ 2	seedream-4-fal閉源ByteDance	1,116	±7	11,870
46	▼ 2	gpt-image-1閉源OpenAI	1,115	±3	264,897
47	▼ 2	recraft-v4閉源Recraft	1,114	±3	114,295
48	▼ 2	seedream-4-high-res-fal閉源ByteDance	1,113	±3	175,361
49	▼ 2	krea-2-turbo開源Krea	1,110	±5	33,237
50	▼ 2	gpt-image-1-mini閉源OpenAI	1,109	±3	163,843
51	▼ 2	krea-2-large閉源Krea	1,108	±5	39,614
52	▼ 2	wan2.7-image-pro閉源Alibaba	1,103	±5	28,729
53	▼ 2	wan2.7-image閉源Alibaba	1,100	±5	29,089
54	▼ 2	mai-image-1閉源Microsoft AI	1,093	±4	94,562
55	▼ 2	z-image-turbo開源Alibaba	1,084	±5	22,251
56	▼ 2	seedream-3閉源ByteDance	1,082	±5	36,920
57	▼ 2	flux-1-kontext-max閉源Black Forest Labs	1,074	±3	65,618
58	▼ 2	flux-2-klein-9b開源Black Forest Labs	1,070	±3	142,642
59	▼ 2	qwen-image-prompt-extend開源Alibaba	1,061	±3	705,193
60	▼ 2	flux-1-kontext-pro閉源Black Forest Labs	1,059	±3	330,851
61	▼ 2	imagen-3.0-generate-002閉源Google	1,058	±3	360,152
62	▼ 2	qwen-image開源Alibaba	1,057	±3	84,717
63	▼ 2	ideogram-v3-quality閉源Ideogram	1,049	±4	115,413
64	▼ 2	photon閉源Luma AI	1,035	±4	127,530
65	▼ 2	p-image閉源	1,034	±4	103,848
66	▼ 2	flux-2-klein-4b開源Black Forest Labs	1,030	±3	144,483
67	▼ 2	runway-gen4閉源Runway	1,025	±5	51,755
68	▼ 2	recraft-v3閉源Recraft	1,021	±4	191,836
69	▼ 2	flux-1.1-pro閉源Black Forest Labs	1,016	±4	70,460
70	▼ 2	lucid-origin閉源Leonardo.Ai	1,013	±3	285,042
71	▼ 2	ideogram-v2閉源Ideogram	1,013	±4	72,090
72	▼ 2	glm-image開源Z.ai	1,010	±9	4,624
73	▼ 2	gemini-2.0-flash-preview-image-generation閉源Google	975	±3	257,085
74	▼ 2	flux-1-dev-fp8開源Black Forest Labs	969	±4	49,221
75	▼ 2	dall-e-3閉源OpenAI	968	±4	239,060
76	▼ 2	flux-1-kontext-dev開源Black Forest Labs	940	±4	215,336
77	▼ 2	stable-diffusion-v35-large開源	938	±5	23,396
78	▼ 2	bagel開源ByteDance	898	±6	12,375

其餘 28 個模型

智力進步

每家廠商當時最強的模型在 Artificial Analysis 智力指數上的分數，按模型釋出的月份畫。

漲得最多 · Anthropic +57.2目前最高 · Anthropic 64.6最高與最低相差 18.1 分時間範圍 2025年11月 – 2026年11月

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