List your company

Mathematical Reasoning on GSM8k (Throughput and Latency)

39.69Throughput (tokens/s)PACER
7.9716.20524.4432.675Feb 1, 2026

Evaluation results

MethodLinks
PACERLLM Family/Size=Deepseek 1.3B/33B2026.0239.692.17
SpeculativeLLM Family/Size=Deepseek 1.3B/33B2026.0235.081.92
AssistLLM Family/Size=Deepseek 1.3B/33B2026.0234.861.91
LookaheadLLM Family/Size=Deepseek 1.3B/33B2026.0233.821.85
PACERLLM Family/Size=Deepseek 6.7B/33B2026.0231.071.88
SpeculativeLLM Family/Size=Deepseek 6.7B/33B2026.0229.361.78
LookaheadLLM Family/Size=Deepseek 6.7B/33B2026.0229.091.76
AssistLLM Family/Size=Deepseek 6.7B/33B2026.0228.81.74
RESTLLM Family/Size=Deepseek 1.3B/33B2026.0223.251.27
RESTLLM Family/Size=Deepseek 6.7B/33B2026.0222.221.34
PACERLLM Family/Size=Llama-2 7B/70B2026.0219.072.08
VanillaLLM Family/Size=Deepseek 1.3B/33B2026.0218.251
SpeculativeLLM Family/Size=Llama-2 7B/70B2026.0217.811.94
VanillaLLM Family/Size=Deepseek 6.7B/33B2026.0216.511
AssistLLM Family/Size=Llama-2 7B/70B2026.0214.021.53
RESTLLM Family/Size=Llama-2 7B/70B2026.0212.661.38
LookaheadLLM Family/Size=Llama-2 7B/70B2026.0212.071.31
VanillaLLM Family/Size=Llama-2 7B/70B2026.029.191