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

Long-context reasoning on LongReason 64K-input 70K context

71.25Accuracy

KVZip

-2.8516.387535.62554.8625Oct 9, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.10
71.25
2025.10
70.25
2025.10
69.75
2025.10
68.5
2025.10
68.25
2025.10
67.25
2025.10
63.25
2025.10
58.25
2025.10
55.5
2025.10
53
2025.10
52.5
2025.10
50.5
2025.10
49.5
2025.10
49.25
2025.10
49
2025.10
48.75
2025.10
48
2025.10
45.25
2025.10
36
2025.10
35.25
2025.10
15
2025.10
4.75
2025.10
3
2025.10
1.5
2025.10
1.25
2025.10
1
2025.10
0.75
2025.10
0.5
2025.10
0.5
2025.10
0.25
2025.10
0
2025.10
0
2025.10
0
2025.10
0