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("MajajHong2015public.IT-reverse_pls")
score = benchmark(my_model)

Benchmark API

Code examples

Model scores

Score Legend Min Alignment Max Alignment

Rank	Model	Score
1	compact_resnet50_robust_IT	.851
2	compact_resnet50_IT	.850
3	compact_vgg19_IT	.841
4	compact_cornets_IT	.835
5	compact_resnet50_robust_V4	.627
6	compact_resnet50_V4	.615
7	compact_vgg19_V4	.614
8	compact_cornets_V4	.613
9	pr1_stl10_disentangled_alpha04	.299
10	alexnet_training_seed_06	.272
11	alexnet_training_seed_03	.231
12	alexnet_training_seed_01	.218
13	gabor_filter_mean_abs_4x4_v2	.217
14	alexnet_training_seed_07	.216
15	predictive_coding_pc	.206
16	gabor_filter_entropy_4x4	.196
17	gabor_filter_entropy_4x4_v2	.196
18	alexnet_training_seed_10	.191
19	gabor_filter_entropy_4x4_v2_forcedCosine	.177
20	mehrer2020_alexnet_training_seed_01_fov4	.173
21	hmax	.172
22	alexnet_training_seed_09	.142
23	alexnet_training_seed_02	.131
24	resnet-50-robust	.125
25	vit_relpos_base_patch32_plus_rpn_256:sw_in1k	.117
26	custom_model_cv_18_dagger_408	.112
27	pnasnet_large_halfPrecision	.111
28	convnext_femto_ols:d1_in1k	.110
29	ReAlnet01_cornet_fov16	.095
30	densenet-121	.093
31	vit_huge_patch14_clip_224:laion2b_ft_in12k_in1k	.092
32	effnetb1_cutmixpatch_SAM_robust32_avge6e8e9e10_manylayers_324x288	.090
33	efficientnet_b0	.087
34	vit_large_patch14_clip_336:openai_ft_in12k_in1k	.087
35	mobilenet_v2_1_4_224	.086
36	swin_small_patch4_window7_224:ms_in22k_ft_in1k	.085
37	vit_base_patch16_clip_224:openai_ft_in1k	.085
38	vit_large_patch14_clip_224:laion2b_ft_in1k	.084
39	vit_large_patch14_clip_224:openai_ft_in1k	.084
40	AlexNet_SIN_fov4	.083
41	resnet34_ecoset_full	.083
42	mehrer2020_alexnet_training_seed_01_fov16	.083
43	AdvProp_efficientnet-b2	.082
44	AT_efficientnet-b2	.082
45	shufflenet_v2_x1_0	.081
46	resnet_SIN_IN_FT_IN	.081
47	resnet18_fair	.080
48	convnext_base:clip_laiona_augreg_ft_in1k_384	.080
49	vit_large_patch14_clip_336:laion2b_ft_in1k	.080
50	vit_large_patch14_clip_224:openai_ft_in12k_in1k	.078
51	effnetb1_cutmix_augmix_sam_e1_5avg_424x377	.076
52	efficientnet_b1_imagenet_full	.075
53	vit_tiny_r_s16_p8_384:augreg_in21k_ft_in1k	.075
54	convnext_tiny_imagenet_full_seed-0	.074
55	ReAlnet01_cornet_fov12	.074
56	convnext_tiny:in12k_ft_in1k	.074
57	VOneCORnet-S	.074
58	efficientnet_b2_imagenet_full	.073
59	efficientnet_b0_imagenet_full	.073
60	ViT_L_32_imagenet1k	.071
61	AdvProp_efficientnet-b4	.071
62	AdvProp_efficientnet-b6	.070
63	convnext_base_imagenet_full_seed-0	.068
64	convnext_small_imagenet_full_seed-0	.068
65	AdvProp_efficientnet-b7	.067
66	AlexNet_SIN_fov16	.066
67	AdvProp_efficientnet-b8	.066
68	convnext_large:fb_in22k_ft_in1k	.065
69	effnetb1_cutmixpatch_augmix_robust32_avge4e7_manylayers_324x288	.065
70	resnet50_robust_l2_eps1	.065
71	convnext_large_mlp:clip_laion2b_augreg_ft_in1k_384	.065
72	grcnn	.065
73	AlexNet_SIN	.063
74	ReAlnet07_cornet	.063
75	effnetb1_272x240	.063
76	convnext_xxlarge:clip_laion2b_soup_ft_in1k	.062
77	ReAlnet09_cornet	.061
78	ReAlnet10_cornet	.061
79	ReAlnet02_cornet	.061
80	ReAlnet05_cornet	.061
81	ReAlnet08_cornet	.061
82	ReAlnet06_cornet	.061
83	ReAlnet01_cornet_fov4	.061
84	ReAlnet01_cornet	.060
85	ReAlnet03_cornet	.059
86	resnet50_imagenet_100_seed-0	.059
87	ReAlnet04_cornet	.058
88	alexnet	.058
89	ijepa_vith14_linhead	.057
90	convnext_large_imagenet_full_seed-0	.057
91	cvt_cvt-w24-384-in22k_finetuned-in1k_4	.055
92	mehrer2020_alexnet_training_seed_01_fov12	.055
93	convnext_tiny_sup	.054
94	cvt_cvt-21-384-in1k_4_LucyV4	.054
95	vit_relpos_base_patch16_clsgap_224:sw_in1k	.053
96	focalnet_tiny_lrf_in1k	.052
97	AlexNet_SIN_fov12	.051
98	cvt_cvt-13-384-in1k_4_LucyV4	.050
99	vit_large_patch14_clip_224:laion2b_ft_in12k_in1k	.050
100	regnety_032	.050
101	deit_small_imagenet_full_seed-0	.049
102	cvt_cvt-13-384-in22k_finetuned-in1k_4_LucyV4	.049
103	convnext_xlarge:fb_in22k_ft_in1k	.049
104	convnext_small_imagenet_100_seed-0	.049
105	vit_base_patch16_clip_224:openai_ft_in12k_in1k	.048
106	deit_large_imagenet_full_seed-0	.048
107	cvt_cvt-21-384-in22k_finetuned-in1k_4_LucyV4	.047
108	mehrer2020_alexnet_training_seed_01	.047
109	mehrer2020_alexnet_training_seed_04	.047
110	seresnext50_32x4d	.045
111	antialiased-r50	.045
112	cvt_cvt-13-224-in1k_4_LucyV4	.043
113	deit_base_imagenet_full_seed-0	.043
114	deit_small_imagenet_100_seed-0	.043
115	resnet50-sup	.042
116	resnet_50_v1	.042
117	resnet_50_v1_spiking	.042
118	vit_large_clip336_openai	.041
119	voneresnet-50-non_stochastic	.040
120	resnet50-SIN	.040
121	convnext_xxlarge_clip	.040
122	cvt_cvt-21-224-in1k_4_LucyV4	.040
123	dinov2_vitl14_lc	.039
124	convnext_base_fb_in1k	.039
125	dinov2_vitl14_lc_res336	.038
126	voneresnet-50	.037
127	dinov2_vitg14_reg_lc	.035
128	resnext101_32x8d_wsl	.035
129	resnet101_imagenet_full	.034
130	dinov2_vitl14_reg_lc_featread	.033
131	dinov2_vitl14_reg_lc	.033
132	resnet152_imagenet_full	.033
133	eicircuit_resnet18_bn_20260919	.031
134	dinov2_vitg14_lc	.031
135	resnext101_32x16d_wsl	.030
136	resnet50-SIN_IN	.030
137	resnext101_32x32d_wsl	.028
138	resnet50-vitoimagevidnet8	.028
139	resnet50-VITO-8deg-cc	.028
140	resnet50_ewconplus_wzc_seed2_2__323	.028
141	resnet50_ewconplus_wzc_seed6_2_323	.028
142	resnet50_ewconplus_wzc	.028
143	resnext101_32x48d_wsl	.027
144	resnet50_ewconplus_wzc_seed5_2__323	.027
145	resnet34_imagenet_full	.026
146	resnet18_ecoset_full	.025
147	mehrer2020_alexnet_training_seed_08	.025
148	resnet18_imagenet_full	.025
149	mehrer2020_alexnet_training_seed_09	.025
150	resnet50_vicreg_linhead	.025
151	mehrer2020_alexnet_training_seed_10	.025
152	resnet50-SIN_IN_IN	.025
153	mehrer2020_alexnet_training_seed_07	.025
154	resnet50_tutorial	.025
155	resnet50_simsiam_linhead	.025
156	resnet50_barlowtwins_linhead	.025
157	mehrer2020_alexnet_training_seed_06	.025
158	Res2Net50_26w_4s	.025
159	resnet50_st_wzc_seed3_2_323	.024
160	resnet50_st_wzc_seed7_2_323	.024
161	mehrer2020_alexnet_training_seed_05	.024
162	resnet50_st_wzc_seed6_2_323	.024
163	resnet50_imagenet_full	.024
164	resnet50-moclr8deg	.023
165	resnet50_ecoset_full	.023
166	mehrer2020_alexnet_training_seed_02	.023
167	resnet152_ecoset_full	.022
168	mehrer2020_alexnet_training_seed_03	.022
169	resnet101_ecoset_full	.022
170	resnet50_simclr_linhead_01norm	.019
171	resnet50_byol_linhead	.018
172	resnet_50_v2	.014
173	vit_huge_clip336_laion
174	eva02_large_in22k

Benchmark bibtex

@article{muzellec_reverse_2026,
      title = {Reverse predictivity for bidirectional comparison of neural networks and biological brains},
      volume = {8},
      issn = {2522-5839},
      url = {https://doi.org/10.1038/s42256-026-01204-0},
      doi = {10.1038/s42256-026-01204-0},
      number = {3},
      journal = {Nature Machine Intelligence},
      author = {Muzellec, Sabine and Kar, Kohitij},
      month = mar,
      year = {2026},
      pages = {474--488},
}

Ceiling

0.82.

Note that scores are relative to this ceiling.

Data: MajajHong2015public.IT

Metric: reverse_pls