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

Long-context understanding on InfiniteBench

0.5Math Score (F)

SinkRouter

-0.0110560.1216220.25430.386978Sep 29, 2025Nov 1, 2025Dec 5, 2025Jan 7, 2026Feb 10, 2026Mar 15, 2026Apr 18, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
0.50.680.27430.190.568--0.22870.5347---10.99490.37660.0004
2026.04
0.4950.6850.28520.190.546--0.23520.5343---10.99490.3774-
2025.09
0.4771------0.3020.37910.30120.15160.7853----
2025.09
0.4371------0.32740.37260.29630.15560.704----
2025.09
0.4229------0.30960.37390.31420.15660.7507----
2025.09
0.4171------0.32990.38330.29750.15360.776----
2025.09
0.4114------0.31980.38220.30570.15650.7787----
2025.09
0.3571------0.2640.34780.290.12570.7359----
2026.04
0.3450.690.2720.170.405--0.2150.4948---0.990.990.3766-0.0395
2025.09
0.3343------0.23350.35380.32090.13650.8484----
2025.09
0.3257------0.22840.34890.31450.13280.8003----
2025.09
0.3229------0.2360.33960.30750.13690.8077----
2025.09
0.32------0.25130.36060.33210.13350.8149----
2026.02
0.25710.65070.28010.170.550.98470.97460.2030.5088-------
2026.02
0.240.64190.27640.190.5520.98470.9780.20560.5086-------
2026.02
0.22860.67250.2760.140.4680.98980.96270.20050.4923-------
2026.02
0.22570.68120.27370.110.5080.99150.96950.2030.4953-------
2026.02
0.22290.67250.2680.1850.5160.97460.9780.19540.5016-------
2026.02
0.21430.65940.27610.1850.4960.97460.9780.19290.497-------
2026.02
0.060.37990.16760.05500.34580.31190.26140.1977-------
2026.02
0.05140.38430.1630.05500.33730.30.26140.1941-------
2026.02
0.04570.35810.17450.0700.41190.65760.26650.248-------
2026.02
0.03140.35370.16710.06500.39320.65930.25630.2408-------
2026.02
0.01430.33190.16870.05500.36610.62710.27160.2293-------
2026.02
0.00860.33190.16810.05500.35760.62880.27410.228-------