List your company

Speculative Decoding on AIME 25

7.08SpeedupTAPS
0.75682.39844.045.6816Jul 2, 2025Aug 26, 2025Oct 20, 2025Dec 15, 2025Feb 8, 2026Apr 4, 2026May 30, 2026

Evaluation results

MethodLinks
TAPSHardware=NVIDIA A40-48GB GPU, Target Model=Qwen3-4B2026.057.08—8.62
TAPSHardware=NVIDIA A40-48GB GPU, Target Model=Qwen3-8B2026.056.64—8.32
DFlashHardware=NVIDIA A40-48GB GPU, Target Model=Qwen3-4B2026.055.68—6.75
DFlashHardware=NVIDIA A40-48GB GPU, Target Model=Qwen3-8B2026.055.62—6.74
TAPSHardware=NVIDIA A800-SXM4-80GB GPU, Target Model=Qwen3-8B2026.054.84—8.1
TAPSHardware=NVIDIA A800-SXM4-80GB GPU, Target Model=Qwen3-4B2026.054.82—8.19
DDTreeHardware=NVIDIA A800-SXM4-80GB GPU, Target Model=Qwen3-4B2026.054.23—9.36
TAPSHardware=NVIDIA A800-SXM4-80GB GPU, Target Model=LLaMA-3.1-8B-Instruct2026.053.98—5.21
DDTreeHardware=NVIDIA A40-48GB GPU, Target Model=Qwen3-4B2026.053.83—9.67
DFlashHardware=NVIDIA A800-SXM4-80GB GPU, Target Model=Qwen3-4B2026.053.73—4.92
DFlashHardware=NVIDIA A800-SXM4-80GB GPU, Target Model=Qwen3-8B2026.053.57—4.73
TAPSHardware=NVIDIA A40-48GB GPU, Target Model=LLaMA-3.1-8B-Instruct2026.053.5—5.16
LogitSpec2025.073.333.76—
DDTreeHardware=NVIDIA A800-SXM4-80GB GPU, Target Model=Qwen3-8B2026.052.96—9.22
DDTreeHardware=NVIDIA A40-48GB GPU, Target Model=Qwen3-8B2026.052.91—9.55
DFlashHardware=NVIDIA A40-48GB GPU, Target Model=LLaMA-3.1-8B-Instruct2026.052.65—3.68
DFlashHardware=NVIDIA A800-SXM4-80GB GPU, Target Model=LLaMA-3.1-8B-Instruct2026.052.56—3.65
DDTreeHardware=NVIDIA A800-SXM4-80GB GPU, Target Model=LLaMA-3.1-8B-Instruct2026.052.25—5.97
EAGLE-3Hardware=NVIDIA A40-48GB GPU, Target Model=LLaMA-3.1-8B-Instruct2026.052.02—2.81
EAGLE-3Hardware=NVIDIA A40-48GB GPU, Target Model=Qwen3-4B2026.051.79—3.05
EAGLE-3Hardware=NVIDIA A40-48GB GPU, Target Model=Qwen3-8B2026.051.79—3
EAGLE-3Hardware=NVIDIA A800-SXM4-80GB GPU, Target Model=Qwen3-4B2026.051.64—2.79
EAGLE-3Hardware=NVIDIA A800-SXM4-80GB GPU, Target Model=LLaMA-3.1-8B-Instruct2026.051.64—2.34
EAGLE-3Hardware=NVIDIA A800-SXM4-80GB GPU, Target Model=Qwen3-8B2026.051.63—2.74
DDTreeHardware=NVIDIA A40-48GB GPU, Target Model=LLaMA-3.1-8B-Instruct2026.051.47—5.91
Vanilla2025.0711—