List your company
ResearchFormerly WizWand

Scientific Reasoning on SuperGPQA

50.1Mean@1Agentic Proposing
16.882425.506234.1342.7538Aug 25, 2025Sep 22, 2025Oct 20, 2025Nov 18, 2025Dec 16, 2025Jan 13, 2026Feb 11, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Agentic ProposingTraining Data=Agentic-Proposer-4B Generated Problems, Training Budget=10,000 trajectories, Optimizer=GRPO2026.0250.1—
Dr.SCI-4B-thinkModel Category=Thinking Models, Thinking Mode=true, Model Scale=4B2026.0245.7—
o1-miniModel Category=Thinking Models, Thinking Mode=true2026.0245.2—
Kimi-K2Model Variant=Base, # Shots=5-shot, # Activated Params=32B, # Total Params=1043B2026.0244.7—
GPT-4oModel Category=Instruct Models2026.0244.4—
Qwen3-14B-MegaScienceModel Category=Instruct Models, Model Scale=14B, Method Variant=MegaScience2026.0244.4—
DeepSeek V3.2Model Variant=Exp Base, # Shots=5-shot, # Activated Params=37B, # Total Params=671B2026.0243.6—
QwQ-32BModel Category=Thinking Models, Thinking Mode=true, Model Scale=32B2026.0243.6—
Dr.SCI-4B-instructModel Category=Instruct Models, Model Scale=4B2026.0243.6—
Qwen3-4B-Instruct-2507Protocol=zero-shot2026.0242.8—
Qwen3-4B thinkingModel Category=Thinking Models, Thinking Mode=true, Model Scale=4B2026.0242.7—
DeepSeek V3.1Model Variant=Base, # Shots=5-shot, # Activated Params=37B, # Total Params=671B2026.0242.3—
R1-0528-Qwen3-8BModel Category=Thinking Models, Thinking Mode=true, Model Scale=8B2026.0242.1—
MiMo-V2 FlashModel Variant=Base, # Shots=5-shot, # Activated Params=15B, # Total Params=309B2026.0241.1—
Step 3.5 FlashModel Variant=Base, # Shots=5-shot, # Activated Params=11B, # Total Params=196B2026.0241—
General-Reasoner-Qw3-14BModel Category=Instruct Models, Model Scale=14B2026.0239.9—
R1-Distill-Qwen-32BModel Category=Thinking Models, Thinking Mode=true, Model Scale=32B2026.0239.3—
Qwen3-8B-MegaScienceModel Category=Instruct Models, Model Scale=8B, Method Variant=MegaScience2026.0238.8—
Qwen3-8B-VeriFreeModel Category=Instruct Models, Model Scale=8B, Method Variant=VeriFree2026.0238—
Qwen3-4B-VeriFreeModel Category=Instruct Models, Model Scale=4B, Method Variant=VeriFree2026.0235.1—
Qwen3-4B-MegaScienceModel Category=Instruct Models, Model Scale=4B, Method Variant=MegaScience2026.0233.1—
General-Reasoner-4BModel Category=Instruct Models, Model Scale=4B2026.0232.5—
Qwen3-4B non-thinkingModel Category=Instruct Models, Thinking Mode=false, Model Scale=4B2026.0232—
SFTBackbone=Qwen2.5-7B-Instruct2025.0829.02—
Qwen3-4B-BaseModel Category=Base, Model Scale=4B2026.0228.5—
PSFTwarm-upBackbone=Qwen2.5-7B-Instruct2025.0828.37—
PSFTBackbone=Qwen2.5-7B-Instruct2025.0828.1—
SFT-KLBackbone=Qwen2.5-7B-Instruct2025.0827.69—
BaseBackbone=Qwen2.5-7B-Instruct2025.0827.59—
PSFTBackbone=Llama3.1-8B-Instruct2025.0825.85—
PSFTwarm-upBackbone=Llama3.1-8B-Instruct2025.0825.26—
SFTBackbone=Llama3.1-8B-Instruct2025.0821.95—
SFT-KLBackbone=Llama3.1-8B-Instruct2025.0819.81—
BaseBackbone=Llama3.1-8B-Instruct2025.0818.16—