List your company
ResearchFormerly WizWand

Advanced Mathematical Reasoning on OlympiadBench (accuracy)

12.8AccuracyTDA-RC
5.9367.7189.511.282Mar 13, 2026
Updated 7mo ago

Evaluation results

MethodLinks
TDA-RCBase LLM=DeepSeek-V32026.0312.8
TDA-RCBase LLM=GPT-4o-mini2026.0312.6
TDA-RCBase LLM=Qwen-Turbo2026.0312.3
Instruction InductionBase LLM=DeepSeek-V32026.0312.2
HoTBase LLM=DeepSeek-V32026.0312
Instruction InductionBase LLM=GPT-4o-mini2026.0311.9
HoTBase LLM=GPT-4o-mini2026.0311.7
Role / Persona PromptingBase LLM=DeepSeek-V32026.0311.5
HoTBase LLM=Qwen-Turbo2026.0311.4
Role / Persona PromptingBase LLM=GPT-4o-mini2026.0311.3
Instruction InductionBase LLM=Qwen-Turbo2026.0311.3
Prompt CanvasBase LLM=DeepSeek-V32026.0311.1
Prompt CanvasBase LLM=GPT-4o-mini2026.0311
Role / Persona PromptingBase LLM=Qwen-Turbo2026.0310.7
Prompt CanvasBase LLM=Qwen-Turbo2026.0310.4
Analogical PromptingBase LLM=DeepSeek-V32026.036.7
Analogical PromptingBase LLM=GPT-4o-mini2026.036.5
Analogical PromptingBase LLM=Qwen-Turbo2026.036.2