List your company

Mathematical Reasoning on AIME 2025 (Accuracy (%))

40.8Accuracy (%)SPIRAL
-1.1129.76920.6531.531Jun 30, 2025

Evaluation results

MethodLinks
SPIRALBackbone=DeepSeek-Distill-Qwen-7B, Training Data=Multi-game2025.0640.8
DeepSeek-Distill-Qwen-7BBackbone=DeepSeek-Distill-Qwen-7B2025.0639.5
SFTBackbone=DeepSeek-Distill-Qwen-7B, Training Data=Multi-game2025.0636.6
SPIRALBackbone=Qwen3-8B, Training Data=Multi-game2025.0616.8
SPIRALBackbone=Qwen3-4B, Training Data=Kuhn Poker2025.0615.6
SFTBackbone=Qwen3-8B, Training Data=Multi-game2025.0615.6
SPIRALBackbone=Qwen3-4B, Training Data=Multi-game2025.0613.3
SFTBackbone=Qwen3-4B, Training Data=Multi-game2025.0611.7
Qwen3-8B-BaseBackbone=Qwen3-8B2025.0611.2
SFTBackbone=Qwen3-4B, Training Data=Kuhn Poker2025.0610.4
Qwen3-4B-BaseBackbone=Qwen3-4B2025.066.2
SPIRALBackbone=Octothinker-8B, Training Data=Multi-game2025.064.8
SFTBackbone=Octothinker-8B, Training Data=Multi-game2025.063.8
SPIRALBackbone=Llama-3.1-8B-Instruct, Training Data=Multi-game2025.061.8
Llama-3.1-8B-InstructBackbone=Llama-3.1-8B-Instruct2025.060.7
SFTBackbone=Llama-3.1-8B-Instruct, Training Data=Multi-game2025.060.7
Octothinker-8B-BaseBackbone=Octothinker-8B2025.060.5