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

LLM Inference Efficiency on AIME

40.6Time To First Token (TTFT)

RelayCaching

22.468144.859267.25389.641Feb 28, 2026
Updated 2mo ago

Evaluation Results

MethodLinks
2026.02
40.62.11.650
2026.02
44.91.9--
2026.02
573.162.860
2026.02
73.24.394.370
2026.02
76.72.35--
2026.02
85.2---
2026.02
104.84.717.980
2026.02
112.52.86--
2026.02
159.33.1--
2026.02
180.2---
2026.02
321.5---
2026.02
493.9---