List your company
ResearchFormerly WizWand

Reasoning on ARC Challenge (P@1, P@32, #Tok)

95.5P@1DeepSeek-V4-pro
19.47639.21358.9578.687Jun 4, 2026Jun 5, 2026Jun 6, 2026Jun 7, 2026Jun 8, 2026
Updated 4mo ago

Evaluation results

MethodLinks
DeepSeek-V4-proModel Scale=Reference, Model Category=Open-source frontier2026.0695.5——
Qwen3-maxModel Scale=Reference, Model Category=Closed-source frontier2026.0693.9——
Qwen3-32BModel Scale=32B, Model Category=Reasoning-specialized models2026.0692——
GPT-5(high)Model Scale=Reference, Model Category=Closed-source frontier2026.0691.8——
GPT-o1-miniModel Scale=Reference, Model Category=Closed-source frontier2026.0690.4——
DeepSeek-R1Model Scale=Reference, Model Category=Open-source frontier2026.0689.6——
DiScOModel Scale=32B, Model Category=Ours2026.0688——
DeepSeek-R1-Distill-Qwen-32BModel Scale=32B, Model Category=Reasoning-specialized models2026.0687.3——
QwQ-32B-previewModel Scale=32B, Model Category=Reasoning-specialized models2026.0687.1——
GPT-o1-previewModel Scale=Reference, Model Category=Closed-source frontier2026.0686.8——
DiScOw/o ITDEModel Scale=32B, Model Category=Ours2026.0686.4——
LLaMA3.1-405B-InstructModel Scale=Reference, Model Category=Open-source frontier2026.0685.6——
Qwen2.5-32B-InstructModel Scale=32B, Model Category=Instruction-tuned models2026.0685.5——
LLaMA3.1-70B-InstructModel Scale=Reference, Model Category=Open-source frontier2026.0685.2——
DiScOModel Scale=7B, Model Category=Ours2026.0684.8——
DeepSeek-V3Model Scale=Reference, Model Category=Open-source frontier2026.0684.7——
Qwen2.5-Math-72B-InstructModel Scale=Reference, Model Category=Open-source frontier2026.0684.4——
ReasonFlux-32BModel Scale=32B, Model Category=Reasoning-specialized models2026.0684.3——
Sky-T1-32B-previewModel Scale=32B, Model Category=Reasoning-specialized models2026.0683.7——
Qwen2.5-Math-7B-InstructModel Scale=7B, Model Category=Math instruction-tuned models2026.0683.6——
DIVERModel Scale=7B, Model Category=Diversity-oriented RL models2026.0683——
DeepSeek-Coder-V2-InstructModel Scale=Reference, Model Category=Open-source frontier2026.0682.5——
DiScOw/o ITDEModel Scale=7B, Model Category=Ours2026.0682.3——
Qwen3-8BModel Scale=7B, Model Category=Reasoning-specialized models2026.0682——
GRPO w/ Clip-higherModel Scale=7B, Model Category=Diversity-oriented RL models2026.0681.6——
rStar-Math-7BModel Scale=7B, Model Category=Reasoning-specialized models2026.0679.3——
Pass@k TrainingModel Scale=7B, Model Category=Diversity-oriented RL models2026.0678.5——
Claude3.5-SonnetModel Scale=Reference, Model Category=Closed-source frontier2026.0676.7——
HRPOBackbone=Qwen2.5-3B-Instruct2026.0675.0697.44277.8
GRPOBackbone=Qwen2.5-3B-Instruct2026.0674.7898.12272.4
CoTBackbone=Qwen2.5-3B-Instruct2026.0674.0598.21106.2
TARPOBackbone=Qwen2.5-3B-Instruct2026.0674.0198.89189.4
ReasonFlux-7BModel Scale=7B, Model Category=Reasoning-specialized models2026.0674——
NuminaMath-72B-CoTModel Scale=Reference, Model Category=Open-source frontier2026.0673.9——
GPT-4oModel Scale=Reference, Model Category=Closed-source frontier2026.0672.8——
Entropy-RoutedBackbone=Qwen2.5-3B-Instruct2026.0671.0996.2595.4
Pure LatentBackbone=Qwen2.5-3B-Instruct2026.0670.5778.1697.1
SuperCorrect-7BModel Scale=7B, Model Category=Reasoning-specialized models2026.0667.3——
Qwen2.5-Math-7BModel Scale=7B, Model Category=Math instruction-tuned models2026.0665.3——
DeepSeek-R1-Distill-Qwen-7BModel Scale=7B, Model Category=Reasoning-specialized models2026.0658.8——
Entropy-RLModel Scale=7B, Model Category=Diversity-oriented RL models2026.0652——
LLaMA3.1-8B-InstructModel Scale=7B, Model Category=General instruction-tuned models2026.0622.7——
Mathstral-7B-v0.1Model Scale=7B, Model Category=General instruction-tuned models2026.0622.4——