Leaderboard Explore Compare Profile Community Blog FAQ Competition 2024 Report Issue

Sample stimuli

How to use

from brainscore_vision import load_benchmark
benchmark = load_benchmark("Zerbe2026_fmri.V4-ood-ridgecv")
score = benchmark(my_model)

Benchmark API

Code examples

Model scores

Score Legend Min Alignment Max Alignment

Rank	Model	Score
1	vit_huge_clip336_laion	.442
2	custom_model_cv_18_dagger_408	.438
3	ijepa_vith14_linhead	.431
4	convnext_large:fb_in22k_ft_in1k	.434
5	convnext_xxlarge:clip_laion2b_soup_ft_in1k	.430
6	cvt_cvt-21-224-in1k_4_LucyV4	.426
7	dinov2_vitg14_reg_lc	.425
8	resnet50_vicreg_linhead	.418
9	convnext_xlarge:fb_in22k_ft_in1k	.423
10	cvt_cvt-13-224-in1k_4_LucyV4	.423
11	resnet50_barlowtwins_linhead	.415
12	ReAlnet07_cornet	.415
13	resnet50_ewconplus_wzc	.420
14	ReAlnet01_cornet	.421
15	ReAlnet03_cornet	.420
16	ReAlnet08_cornet	.413
17	ReAlnet01_cornet_fov12	.413
18	ReAlnet09_cornet	.413
19	ReAlnet04_cornet	.419
20	dinov2_vitg14_lc	.418
21	vit_large_clip336_openai	.418
22	pnasnet_large_halfPrecision	.417
23	ReAlnet05_cornet	.411
24	ReAlnet06_cornet	.411
25	ReAlnet10_cornet	.411
26	ReAlnet02_cornet	.409
27	dinov2_vitl14_lc_res336	.415
28	cvt_cvt-w24-384-in22k_finetuned-in1k_4	.412
29	resnet50_simclr_linhead_01norm	.406
30	dinov2_vitl14_reg_lc_featread	.410
31	dinov2_vitl14_reg_lc	.410
32	resnet101_imagenet_full	.403
33	cvt_cvt-13-384-in1k_4_LucyV4	.409
34	resnet152_imagenet_full	.407
35	vit_tiny_r_s16_p8_384:augreg_in21k_ft_in1k	.407
36	ReAlnet01_cornet_fov16	.400
37	resnet50_simsiam_linhead	.399
38	regnety_032	.400
39	convnext_tiny:in12k_ft_in1k	.404
40	cvt_cvt-21-384-in22k_finetuned-in1k_4_LucyV4	.404
41	resnet50-moclr8deg	.403
42	efficientnet_b2_imagenet_full	.403
43	cvt_cvt-13-384-in22k_finetuned-in1k_4_LucyV4	.401
44	deit_small_imagenet_full_seed-0	.394
45	voneresnet-50	.399
46	shufflenet_v2_x1_0	.398
47	resnet101_ecoset_full	.393
48	VOneCORnet-S	.399
49	resnet34_ecoset_full	.391
50	effnetb1_272x240	.397
51	seresnext50_32x4d	.396
52	resnet152_ecoset_full	.389
53	efficientnet_b0	.394
54	efficientnet_b1_imagenet_full	.388
55	vit_large_patch14_clip_224:openai_ft_in12k_in1k	.394
56	convnext_xxlarge_clip	.392
57	AdvProp_efficientnet-b4	.392
58	antialiased-r50	.391
59	resnext101_32x48d_wsl	.392
60	vit_large_patch14_clip_224:openai_ft_in1k	.392
61	AdvProp_efficientnet-b7	.390
62	Res2Net50_26w_4s	.389
63	focalnet_tiny_lrf_in1k	.388
64	cvt_cvt-21-384-in1k_4_LucyV4	.387
65	deit_large_imagenet_full_seed-0	.381
66	convnext_tiny_sup	.386
67	effnetb1_cutmix_augmix_sam_e1_5avg_424x377	.387
68	resnet_SIN_IN_FT_IN	.385
69	resnet50-SIN_IN_IN	.385
70	vit_base_patch16_clip_224:openai_ft_in12k_in1k	.385
71	efficientnet_b0_imagenet_full	.380
72	resnext101_32x16d_wsl	.385
73	mobilenet_v2_1_4_224	.384
74	resnext101_32x32d_wsl	.384
75	vit_relpos_base_patch16_clsgap_224:sw_in1k	.384
76	resnext101_32x8d_wsl	.384
77	convnext_femto_ols:d1_in1k	.383
78	resnet50-vitoimagevidnet8	.383
79	resnet50-VITO-8deg-cc	.383
80	AdvProp_efficientnet-b6	.383
81	AdvProp_efficientnet-b8	.382
82	eicircuit_resnet18_bn_20260919	.375
83	resnet_50_v1_spiking	.379
84	resnet50_tutorial	.379
85	resnet50-sup	.379
86	resnet_50_v1	.379
87	resnet_50_v2	.378
88	resnet50-SIN_IN	.377
89	resnet50_imagenet_full	.377
90	vit_base_patch16_clip_224:openai_ft_in1k	.376
91	convnext_tiny_imagenet_full_seed-0	.375
92	resnet50-SIN	.375
93	AdvProp_efficientnet-b2	.373
94	AT_efficientnet-b2	.373
95	grcnn	.373
96	convnext_large_imagenet_full_seed-0	.373
97	vit_large_patch14_clip_336:openai_ft_in12k_in1k	.373
98	swin_small_patch4_window7_224:ms_in22k_ft_in1k	.369
99	convnext_base_imagenet_full_seed-0	.364
100	densenet-121	.363
101	effnetb1_cutmixpatch_augmix_robust32_avge4e7_manylayers_324x288	.361
102	dinov2_vitl14_lc	.360
103	effnetb1_cutmixpatch_SAM_robust32_avge6e8e9e10_manylayers_324x288	.357
104	voneresnet-50-non_stochastic	.356
105	convnext_large_mlp:clip_laion2b_augreg_ft_in1k_384	.355
106	deit_base_imagenet_full_seed-0	.349
107	convnext_base:clip_laiona_augreg_ft_in1k_384	.354
108	vit_large_patch14_clip_224:laion2b_ft_in12k_in1k	.354
109	vit_huge_patch14_clip_224:laion2b_ft_in12k_in1k	.353
110	convnext_small_imagenet_full_seed-0	.352
111	eva02_large_in22k	.347
112	vit_relpos_base_patch32_plus_rpn_256:sw_in1k	.347
113	convnext_small_imagenet_100_seed-0	.328
114	ReAlnet01_cornet_fov4	.331
115	resnet50_robust_l2_eps1	.325
116	gabor_filter_entropy_4x4_v2	.319
117	alexnet	.320
118	gabor_filter_entropy_4x4_v2_forcedCosine	.317
119	mehrer2020_alexnet_training_seed_09	.315
120	mehrer2020_alexnet_training_seed_08	.311
121	deit_small_imagenet_100_seed-0	.309
122	mehrer2020_alexnet_training_seed_07	.308
123	gabor_filter_entropy_4x4	.307
124	resnet50_ecoset_full	.305
125	resnet-50-robust	.309
126	vit_large_patch14_clip_336:laion2b_ft_in1k	.310
127	mehrer2020_alexnet_training_seed_06	.305
128	vit_large_patch14_clip_224:laion2b_ft_in1k	.305
129	resnet50_imagenet_100_seed-0	.300
130	mehrer2020_alexnet_training_seed_04	.298
131	AlexNet_SIN	.301
132	mehrer2020_alexnet_training_seed_05	.296
133	mehrer2020_alexnet_training_seed_03	.295
134	mehrer2020_alexnet_training_seed_01	.293
135	mehrer2020_alexnet_training_seed_02	.289
136	mehrer2020_alexnet_training_seed_01_fov12	.266
137	mehrer2020_alexnet_training_seed_10	.265
138	ViT_L_32_imagenet1k	.267
139	mehrer2020_alexnet_training_seed_01_fov16	.256
140	AlexNet_SIN_fov12	.258
141	hmax	.211
142	AlexNet_SIN_fov16	.202
143	gabor_filter_mean_abs_4x4_v2	.179
144	AlexNet_SIN_fov4	.177
145	alexnet_training_seed_09	.151
146	alexnet_training_seed_02	.146
147	alexnet_training_seed_06	.146
148	alexnet_training_seed_07	.141
149	alexnet_training_seed_10	.137
150	alexnet_training_seed_03	.136
151	alexnet_training_seed_01	.130
152	compact_resnet50_robust_V4	.116
153	resnet50_byol_linhead
154	mehrer2020_alexnet_training_seed_01_fov4
155	resnet34_imagenet_full
156	resnet18_imagenet_full
157	resnet18_ecoset_full

Benchmark bibtex

@inproceedings{zerbe_laion-fmri_2026,
      title = {{LAION}-{fMRI}: A densely sampled 7T-fMRI dataset providing broad coverage of natural image diversity},
      author = {Zerbe, Josefine and Roth, Johannes and Mell, Maggie Mae and Herholz, Peer and Knapen, Tomas and Hebart, Martin N.},
      year = {2026},
      booktitle = {Vision Sciences Society Annual Meeting},
  }

Ceiling

0.79.

Note that scores are relative to this ceiling.

Data: Zerbe2026_fmri.V4-ood

Metric: ridgecv