List your company
ResearchFormerly WizWand

Mathematical Reasoning on AIME 2024 (Token Usage and Cost)

25,573.15Input TokensBoT
-740.88846,090.640812,922.1719,753.6992Sep 26, 2025
Updated 7mo ago

Evaluation results

MethodLinks
BoTLLM=Qwen3-8B2025.0925,573.1524,4201.51
BoTLLM=Qwen3-4B2025.0922,015.5621,158.810.79
BoTLLM=Qwen3-1.7B2025.0915,296.4826,651.480.95
BoTLLM=Qwen3-0.6B2025.096,530.729,351.31.02
CoT-SCLLM=Qwen3-0.6B2025.0990046,658.931.59
CoT-SCLLM=Qwen3-8B2025.09899.7840,227.72.29
CoT-SCLLM=Qwen3-14B2025.09891.4136,064.264.1
CoT-SCLLM=Qwen3-4B2025.09889.4837,251.441.27
CoT-SCLLM=Qwen3-1.7B2025.09598.7429,708.441.01
RAGLLM=Qwen3-0.6B2025.0938612,147.040.41
RAGLLM=Qwen3-1.7B2025.0938611,844.670.4
RAGLLM=Qwen3-4B2025.093869,949.70.34
RAGLLM=Qwen3-8B2025.0938611,897.480.68
RoT + TILLM=Qwen3-0.6B2025.09343.593,975.110.14
RoT + TILLM=Qwen3-1.7B2025.09343.598,830.370.3
RoT + TILLM=Qwen3-4B2025.09343.599,1380.31
RoT + TILLM=Qwen3-8B2025.09343.5910,586.440.6
RoT + TILLM=Qwen3-14B2025.09343.5910,857.441.23
CoTLLM=Qwen3-0.6B2025.09279.5913,253.440.45
CoTLLM=Qwen3-1.7B2025.09279.5911,246.30.38
CoTLLM=Qwen3-4B2025.09279.599,796.560.33
CoTLLM=Qwen3-8B2025.09279.5910,952.960.62
CoTLLM=Qwen3-14B2025.09279.599,405.631.07
RoTLLM=Qwen3-0.6B2025.09273.4411,678.850.4
RoTLLM=Qwen3-1.7B2025.09273.4411,163.440.38
RoTLLM=Qwen3-4B2025.09273.4410,347.890.35
RoTLLM=Qwen3-8B2025.09273.4411,288.480.64
RoTLLM=Qwen3-14B2025.09273.4410,8051.23
RAGLLM=Qwen3-14B2025.09271.1910,710.191.22