Leaderboard

An overview of pathology foundation models performance across multiple benchmarks

(48/48 models)
Model
ViT-H/14 (632M)
5/13
1/15
5/25
2/21
2/5
3/23
12/17
5/16
4/11
2/13
1/22
2/23
12
ViT-g/14 (1.1B)
4/13
4/15
2/25
3/21
5/5
9/23
1/17
4/16
1/11
5/13
3/22
9/23
12
UNI
ViT-L/16 (307M)
10/13
5/15
11/25
6/21
-
16/23
7/17
9/16
3/11
6/13
2/22
6/23
11
ViT-g/14 (1.1B)
9/13
9/15
10/25
8/21
-
14/23
10/17
7/16
2/11
3/13
4/22
10/23
11
ViT-H/14 (682M)
6/13
2/15
3/25
4/21
3/5
11/23
3/17
8/16
--
5/22
1/23
10
ViT-B/16 (86M)
-
11/15
18/25
12/21
-
21/23
14/17
13/16
5/11
12/13
13/22
19/23
10
ViT-B/16 (86M)
8/13
-
16/25
9/21
-
10/23
11/17
3/16
-
1/13
12/22
16/23
9
ViT-L/14 (307M)
-
13/15
14/25
10/21
-
22/23
13/17
15/16
6/11
-
15/22
20/23
9
ViT-H/14 (632M)
7/13
-
4/25
14/21
-
6/23
4/17
-
9/11
9/13
7/22
13/23
9
ViT-L/14 (307M)
-
8/15
9/25
18/21
--
9/17
16/16
-
7/13
10/22
11/23
8
Swin-T/4 (28M)
11/13
-
23/25
15/21
-
18/23
--
10/11
8/13
22/22
-
7
ViT-g/14 (1.1B)
1/13
-
1/25
1/21
-
8/23
2/17
----
7/23
6
ViT-L/16 (307M)
3/13
-
13/25
11/21
-
4/23
8/17
----
18/23
6
ViT-B/14 (86M)
2/13
-
6/25
--
7/23
-
1/16
--
6/22
4/23
6
V-FFN (202M)
--
24/25
13/21
-
12/23
17/17
----
21/23
5
PLIP
ViT-B/32 (88M)
---
19/21
---
12/16
-
13/13
19/22
22/23
5
ViT-g/14 (1.1B)
-
3/15
7/25
---
16/17
6/16
---
5/23
5
ViT-B/8 (86M)
-
6/15
15/25
--
15/23
-
10/16
--
14/22
-
5
ViT-L/14 (307M)
-
7/15
20/25
---
5/17
14/16
-
11/13
--
5
ViT-L/14 (303M)
--
12/25
7/21
---
11/16
--
9/22
-
4
ResNet-50 (25M)
13/13
-
25/25
20/21
----
11/11
---
4
ViT-B/16 (86M)
-
10/15
19/25
--------
14/23
3
ViT-S/16 (22M)
-
14/15
22/25
--------
17/23
3
ViT-S/16 (22M)
-
15/15
---
19/23
----
11/22
-
3
ViT-S/8 (22M)
--
17/25
---
6/17
--
4/13
--
3
ViT-g/16 (1.1B)
-----
2/23
-
2/16
---
3/23
3
ViT-S/8 (22M)
-
12/15
21/25
---------
2
ViT-S/16 (21M)
-----
13/23
----
17/22
-
2
ViT-B/32 (86M)
----------
21/22
23/23
2
Atlas 2
ViT-G/8 (2.0B)
----
1/5
1/23
------
2
Atlas
ViT-H/14 (632M)
----
4/5
5/23
------
2
ViT-B/14 (86M)
---------
10/13
-
12/23
2
ViT-g/14 (1.1B)
--
8/25
--------
15/23
2
RetCCL
ResNet-50 (23M)
12/13
----
17/23
------
2
ViT-L/14 (307M)
---
5/21
--------
1
ViT-L/14 (307M)
---
16/21
--------
1
ViT-L/16 (306M)
---
21/21
--------
1
ViT-B/16 (85M)
----------
16/22
-
1
KEEP
ViT-L/16 (307M)
-----------
8/23
1
CHIEF
Swin-T (CTransPath) (28M)
---
17/21
--------
1
Ciga
ResNet-18 (11.7M)
-----
23/23
------
1
RudolfV
ViT-L/14 (307M)
-----
20/23
------
1
ViT-S/16 (22M)
--------
8/11
---
1
ViT-B/16 (85M)
--------
7/11
---
1
ViT-B/16 (86M)
----------
20/22
-
1
ViT-B/16 (86M)
----------
18/22
-
1
ViT-B/16 (86M)
----------
8/22
-
1
ViT-L/16 (307M)
------
15/17
-----
1

Direct comparison

Compare pathology foundation models on a common set of benchmarks

Robustness:
Performance:
Models:

Scroll to zoom, drag to pan. Double-click to reset.

13 models with data for all selected benchmarks.

Scaling laws

Explore how model size correlates with performance or robustness across benchmarks

Benchmarks:
Models:

Scroll to zoom, drag to pan. Double-click to reset.

Linear regression: R² = 0.210

13 models with data for all selected benchmarks.