Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Multimodal Mathematical Reasoning on DynaMath (accuracy)

69.2Accuracy (DynaMath)

Metis

18.34431.54744.7557.953Oct 27, 2025Nov 30, 2025Jan 3, 2026Feb 6, 2026Mar 12, 2026Apr 15, 2026May 19, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
69.2-
2026.04
65.5-
2026.05
65.37-
2026.05
65.07-
2026.05
64.59-
2026.05
63.97-
2025.11
61.9-
2026.05
61.8-
2026.05
61.58-
2025.11
57.2-
2026.04
57.2-
2025.11
55-
2026.04
55-
2025.11
53.3-
2026.04
53.3-
2026.05
51.44-
2025.10
50.7-
2026.05
50.36-
2026.05
50.36-
2026.05
50.08-
2025.10
47.5-
2025.10
47.2-
2026.05
46.85-
2026.05
46.65-
2025.10
42.7-
2025.10
42.7-
2026.03
25.2-
2026.03
24.2-
2026.03
23.8-
2026.03
23.4-
2026.03
23.3-
2026.03
23.2-
2026.03
23-
2026.03
22.9-
2026.03
22.6-
2026.03
22.1-
2026.03
21.9-
2026.03
21.8-
2026.03
21.8-
2026.03
21.7-
2026.03
21.4-
2026.03
21.3-
2026.03
21.2-
2026.03
21-
2026.03
20.3-
2026.04
-34.5
2026.04
-36.6
2026.04
-40.7
2026.04
-9.4
2026.04
-9.8
2026.04
-26.7
2026.04
-18.1
2026.04
-15.6
2026.04
-11.4
2026.04
-21
2026.04
-23.7
2026.04
-21
2026.04
-26.8
2026.04
-29.2