List your company

Inference Latency on MMLU-Pro

3.3Latency (s)T2
-6.60860.271127.15194.029Jan 29, 2026

Evaluation results

MethodLinks
T2Backbone=Qwen2.5-7B2026.013.3
T2Backbone=Llama3.1-8B2026.013.7
Std-PTBackbone=Qwen2.5-7B2026.018.5
Std-PTBackbone=Llama3.1-8B2026.018.9
T2Backbone=Qwen3-4B2026.0112.3
ESCBackbone=Qwen2.5-7B2026.0115.1
ACBackbone=Qwen2.5-7B2026.0117.3
ESCBackbone=Llama3.1-8B2026.0119.2
DSCBackbone=Llama3.1-8B2026.0125.2
DeepConfBackbone=Qwen2.5-7B2026.0127.3
ACBackbone=Llama3.1-8B2026.0128.2
Std-PTBackbone=Qwen3-4B2026.0130.7
T2Backbone=Deepseek-R1-Distill-Qwen-7B2026.0131.7
DSCBackbone=Qwen2.5-7B2026.0131.8
DeepConfBackbone=Llama3.1-8B2026.0137
Std-PTBackbone=Deepseek-R1-Distill-Qwen-7B2026.0164.7
ESCBackbone=Qwen3-4B2026.0168.5
ACBackbone=Qwen3-4B2026.0172.5
DSCBackbone=Qwen3-4B2026.0191.2
DSCBackbone=Deepseek-R1-Distill-Qwen-7B2026.01115.8
ESCBackbone=Deepseek-R1-Distill-Qwen-7B2026.01124.2
ACBackbone=Deepseek-R1-Distill-Qwen-7B2026.01150.1
DeepConfBackbone=Qwen3-4B2026.01153.2
DeepConfBackbone=Deepseek-R1-Distill-Qwen-7B2026.01251