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("Cowley2026.V4-pls")
score = benchmark(my_model)

Benchmark API

Code examples

Model scores

Score Legend Min Alignment Max Alignment

Rank	Model	Score
1	custom_model_cv_18_dagger_408	.356
2	voneresnet-50	.334
3	alexnet	.332
4	convnext_tiny:in12k_ft_in1k	.330
5	shufflenet_v2_x1_0	.329
6	AT_efficientnet-b2	.329
7	AdvProp_efficientnet-b2	.329
8	AlexNet_SIN_fov12	.328
9	convnext_xxlarge:clip_laion2b_soup_ft_in1k	.328
10	resnet50-SIN	.327
11	resnet50-moclr8deg	.321
12	vit_relpos_base_patch16_clsgap_224:sw_in1k	.322
13	AdvProp_efficientnet-b4	.322
14	AdvProp_efficientnet-b8	.318
15	cvt_cvt-13-224-in1k_4_LucyV4	.316
16	vit_large_patch14_clip_336:laion2b_ft_in1k	.315
17	efficientnet_b0	.314
18	AlexNet_SIN	.315
19	ReAlnet04_cornet	.312
20	cvt_cvt-21-224-in1k_4_LucyV4	.313
21	ReAlnet01_cornet	.311
22	ReAlnet03_cornet	.311
23	cvt_cvt-13-384-in1k_4_LucyV4	.311
24	resnet152_imagenet_full	.309
25	VOneCORnet-S	.307
26	resnet50-VITO-8deg-cc	.305
27	resnet50-vitoimagevidnet8	.305
28	efficientnet_b2_imagenet_full	.303
29	cvt_cvt-21-384-in1k_4_LucyV4	.303
30	resnet50_ewconplus_wzc	.301
31	focalnet_tiny_lrf_in1k	.298
32	cvt_cvt-21-384-in22k_finetuned-in1k_4_LucyV4	.297
33	mobilenet_v2_1_4_224	.296
34	vit_large_patch14_clip_224:openai_ft_in1k	.295
35	convnext_tiny_sup	.295
36	cvt_cvt-13-384-in22k_finetuned-in1k_4_LucyV4	.291
37	effnetb1_272x240	.289
38	vit_base_patch16_clip_224:openai_ft_in12k_in1k	.288
39	deit_small_imagenet_full_seed-0	.287
40	deit_base_imagenet_full_seed-0	.284
41	resnet_SIN_IN_FT_IN	.283
42	resnet50-SIN_IN_IN	.283
43	convnext_large:fb_in22k_ft_in1k	.283
44	resnet50-SIN_IN	.283
45	vit_large_patch14_clip_336:openai_ft_in12k_in1k	.282
46	resnet_50_v1	.280
47	resnet50_tutorial	.280
48	resnet50-sup	.280
49	convnext_femto_ols:d1_in1k	.281
50	convnext_tiny_imagenet_full_seed-0	.281
51	convnext_xlarge:fb_in22k_ft_in1k	.276
52	resnet50_imagenet_full	.276
53	vit_tiny_r_s16_p8_384:augreg_in21k_ft_in1k	.274
54	pnasnet_large_halfPrecision	.274
55	resnet-50-robust	.273
56	Res2Net50_26w_4s	.271
57	deit_large_imagenet_full_seed-0	.272
58	vit_large_patch14_clip_224:openai_ft_in12k_in1k	.271
59	effnetb1_cutmixpatch_SAM_robust32_avge6e8e9e10_manylayers_324x288	.271
60	effnetb1_cutmixpatch_augmix_robust32_avge4e7_manylayers_324x288	.270
61	convnext_large_imagenet_full_seed-0	.268
62	convnext_base_imagenet_full_seed-0	.266
63	vit_large_patch14_clip_224:laion2b_ft_in1k	.265
64	convnext_small_imagenet_full_seed-0	.264
65	convnext_large_mlp:clip_laion2b_augreg_ft_in1k_384	.263
66	resnet_50_v2	.263
67	antialiased-r50	.262
68	resnext101_32x32d_wsl	.262
69	resnext101_32x8d_wsl	.261
70	resnext101_32x48d_wsl	.261
71	vit_base_patch16_clip_224:openai_ft_in1k	.261
72	convnext_base:clip_laiona_augreg_ft_in1k_384	.259
73	resnet50_robust_l2_eps1	.258
74	voneresnet-50-non_stochastic	.256
75	AdvProp_efficientnet-b7	.256
76	vit_relpos_base_patch32_plus_rpn_256:sw_in1k	.254
77	AdvProp_efficientnet-b6	.250
78	effnetb1_cutmix_augmix_sam_e1_5avg_424x377	.249
79	vit_huge_patch14_clip_224:laion2b_ft_in12k_in1k	.246
80	vit_large_patch14_clip_224:laion2b_ft_in12k_in1k	.245
81	eicircuit_resnet18_bn_20260919	.239
82	swin_small_patch4_window7_224:ms_in22k_ft_in1k	.237
83	grcnn	.231
84	densenet-121	.220
85	ViT_L_32_imagenet1k	.217
86	AlexNet_SIN_fov16	.215
87	ReAlnet01_cornet_fov4	.159
88	alexnet_training_seed_06	.132
89	alexnet_training_seed_02	.131
90	alexnet_training_seed_09	.127
91	alexnet_training_seed_01	.121
92	alexnet_training_seed_10	.119
93	alexnet_training_seed_03	.119
94	alexnet_training_seed_07	.116
95	hmax	.090
96	compact_resnet50_robust_V4	.047
97	resnext101_32x16d_wsl
98	resnet_50_v1_spiking
99	cvt_cvt-w24-384-in22k_finetuned-in1k_4

Benchmark bibtex

@article{cowley2026compact,
    title={Compact deep neural network models of the visual cortex},
    author={Cowley, Benjamin R and Stan, Patricia L and Pillow, Jonathan W and Smith, Matthew A},
    journal={Nature},
    volume={652},
    number={8111},
    pages={947--954},
    year={2026},
    publisher={Nature Publishing Group},
    doi={10.1038/s41586-026-10150-1},
    url={https://doi.org/10.1038/s41586-026-10150-1}}

Ceiling

0.80.

Note that scores are relative to this ceiling.

Data: Cowley2026.V4

Metric: pls