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.V4-reverse_pls")
score = benchmark(my_model)

Benchmark API

Code examples

Model scores

Score Legend Min Alignment Max Alignment

Rank	Model	Score
1	compact_resnet50_V4	.883
2	compact_resnet50_robust_V4	.882
3	compact_vgg19_V4	.867
4	compact_cornets_V4	.833
5	compact_cornets_IT	.528
6	compact_vgg19_IT	.520
7	predictive_coding_pc	.160
8	pr1_stl10_disentangled_alpha04	.150
9	alexnet_training_seed_03	.063
10	hmax	.061
11	AlexNet_SIN_fov16	.059
12	resnet-50-robust	.058
13	custom_model_cv_18_dagger_408	.054
14	alexnet_training_seed_07	.051
15	alexnet_training_seed_01	.051
16	alexnet_training_seed_09	.050
17	alexnet_training_seed_06	.049
18	alexnet_training_seed_02	.048
19	vit_large_patch14_clip_336:laion2b_ft_in1k	.046
20	alexnet_training_seed_10	.045
21	dinov2_vitl14_lc	.044
22	mobilenet_v2_1_4_224	.042
23	AlexNet_SIN_fov12	.040
24	vit_relpos_base_patch32_plus_rpn_256:sw_in1k	.036
25	vit_base_patch16_clip_224:openai_ft_in12k_in1k	.035
26	vit_large_patch14_clip_224:laion2b_ft_in12k_in1k	.033
27	ReAlnet01_cornet_fov4	.031
28	vit_large_patch14_clip_224:laion2b_ft_in1k	.031
29	eva02_large_in22k	.030
30	VOneCORnet-S	.030
31	vit_large_patch14_clip_224:openai_ft_in1k	.029
32	densenet-121	.029
33	AdvProp_efficientnet-b6	.028
34	efficientnet_b0	.027
35	ReAlnet03_cornet	.026
36	AdvProp_efficientnet-b2	.026
37	AT_efficientnet-b2	.026
38	grcnn	.026
39	ReAlnet01_cornet	.026
40	resnet50_robust_l2_eps1	.026
41	vit_tiny_r_s16_p8_384:augreg_in21k_ft_in1k	.026
42	ReAlnet04_cornet	.025
43	effnetb1_cutmixpatch_SAM_robust32_avge6e8e9e10_manylayers_324x288	.025
44	ViT_L_32_imagenet1k	.025
45	deit_large_imagenet_full_seed-0	.025
46	efficientnet_b2_imagenet_full	.024
47	AdvProp_efficientnet-b7	.024
48	vit_base_patch16_clip_224:openai_ft_in1k	.023
49	AdvProp_efficientnet-b8	.023
50	swin_small_patch4_window7_224:ms_in22k_ft_in1k	.023
51	effnetb1_cutmixpatch_augmix_robust32_avge4e7_manylayers_324x288	.023
52	AdvProp_efficientnet-b4	.022
53	effnetb1_cutmix_augmix_sam_e1_5avg_424x377	.022
54	cvt_cvt-21-384-in1k_4_LucyV4	.022
55	convnext_femto_ols:d1_in1k	.022
56	deit_base_imagenet_full_seed-0	.021
57	cvt_cvt-13-384-in1k_4_LucyV4	.021
58	shufflenet_v2_x1_0	.020
59	convnext_xxlarge_clip	.020
60	focalnet_tiny_lrf_in1k	.020
61	deit_small_imagenet_full_seed-0	.019
62	convnext_tiny_sup	.019
63	vit_large_clip336_openai	.019
64	resnet50-SIN	.019
65	resnet152_imagenet_full	.019
66	convnext_tiny:in12k_ft_in1k	.018
67	cvt_cvt-21-384-in22k_finetuned-in1k_4_LucyV4	.018
68	vit_huge_clip336_laion	.018
69	cvt_cvt-21-224-in1k_4_LucyV4	.018
70	convnext_large_mlp:clip_laion2b_augreg_ft_in1k_384	.017
71	cvt_cvt-13-224-in1k_4_LucyV4	.017
72	voneresnet-50-non_stochastic	.017
73	convnext_xlarge:fb_in22k_ft_in1k	.017
74	resnet50-vitoimagevidnet8	.017
75	resnet50-VITO-8deg-cc	.017
76	resnet50_ewconplus_wzc_seed2_2__323	.017
77	AlexNet_SIN	.016
78	effnetb1_272x240	.016
79	resnet50_ewconplus_wzc	.016
80	convnext_base:clip_laiona_augreg_ft_in1k_384	.016
81	vit_relpos_base_patch16_clsgap_224:sw_in1k	.016
82	convnext_large_imagenet_full_seed-0	.015
83	resnet50_st_wzc_seed3_2_323	.015
84	Res2Net50_26w_4s	.015
85	resnet50-moclr8deg	.015
86	resnet50_imagenet_full	.015
87	dinov2_vitg14_lc	.015
88	resnet50_ewconplus_wzc_seed5_2__323	.015
89	voneresnet-50	.014
90	resnet50_st_wzc_seed7_2_323	.014
91	dinov2_vitg14_reg_lc	.014
92	resnet50_ewconplus_wzc_seed6_2_323	.014
93	dinov2_vitl14_reg_lc	.013
94	resnet18_fair	.013
95	antialiased-r50	.013
96	compact_resnet50_robust_IT	.012
97	compact_resnet50_IT	.012
98	alexnet	.011
99	convnext_base_fb_in1k	.011
100	resnext101_32x16d_wsl	.011
101	resnet50-SIN_IN	.011
102	resnet50_tutorial	.011
103	resnet_50_v1	.011
104	resnet50-sup	.011
105	resnet_50_v1_spiking	.011
106	resnext101_32x48d_wsl	.010
107	resnet50-SIN_IN_IN	.010
108	resnet_SIN_IN_FT_IN	.010
109	resnet50_st_wzc_seed6_2_323	.010
110	convnext_tiny_imagenet_full_seed-0	.009
111	convnext_small_imagenet_full_seed-0	.008
112	convnext_base_imagenet_full_seed-0	.008
113	cvt_cvt-13-384-in22k_finetuned-in1k_4_LucyV4	.007
114	resnet_50_v2	.006
115	dinov2_vitl14_reg_lc_featread
116	dinov2_vitl14_lc_res336
117	resnext101_32x8d_wsl
118	pnasnet_large_halfPrecision
119	vit_large_patch14_clip_336:openai_ft_in12k_in1k
120	vit_large_patch14_clip_224:openai_ft_in12k_in1k
121	vit_huge_patch14_clip_224:laion2b_ft_in12k_in1k
122	resnext101_32x32d_wsl
123	convnext_xxlarge:clip_laion2b_soup_ft_in1k
124	convnext_large:fb_in22k_ft_in1k
125	cvt_cvt-w24-384-in22k_finetuned-in1k_4

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

0.88.

Note that scores are relative to this ceiling.

Data: MajajHong2015public.V4

Metric: reverse_pls