List your company
ResearchFormerly WizWand

Planning Accuracy on Blocksworld (test)

98.2AccuracyDeepSeek-R1
-3.30423.04849.475.752May 22, 2026May 29, 2026Jun 6, 2026Jun 13, 2026Jun 21, 2026Jun 28, 2026Jul 6, 2026
Updated 3mo ago

Evaluation results

MethodLinks
DeepSeek-R12026.0798.2
o1-preview2026.0797.8
PT-SFTBase model=Mistral2026.0597
PT-SFTBase model=Qwen2.52026.0596
PT-SFTBase model=GPT-OSS2026.0593
QwQ-32B2026.0788.8
HyperGuideBase model=Qwen2.52026.0587
HyperGuideBase model=GPT-OSS2026.0583
SoftCoTBase model=Qwen2.52026.0582
OVMBase model=Qwen2.52026.0581.4
SoftCoTBase model=GPT-OSS2026.0579
OVMBase model=GPT-OSS2026.0577.4
Qwen3-8BTraining=CARL2026.0777.2
HyperGuideBase model=Mistral2026.0576
Qwen3-8BTraining=RFT2026.0773.8
SoftCoTBase model=Mistral2026.0571
OVMBase model=Mistral2026.0562
Self-ConsistencyBase model=Qwen2.52026.0560
Tree of ThoughtsBase model=Qwen2.52026.0558
Few-shotBase model=Mistral2026.0557
Tree of ThoughtsBase model=Mistral2026.0555
Self-ConsistencyBase model=Mistral2026.0553
DeepSeek-R1-Distill-Llama-8BTraining=CARL2026.0752.6
DeepSeek-V32026.0744.8
GPT-4o2026.0742.4
DeepSeek-R1-Distill-Llama-8BTraining=RFT2026.0742
Few-shotBase model=Qwen2.52026.0541
Tree of ThoughtsBase model=GPT-OSS2026.0537
Qwen3-8B2026.0731.2
Llama-3.1-70B-Instruct2026.0721.6
Qwen2.5-72B-Instruct2026.0713.8
Few-shotBase model=GPT-OSS2026.059
Self-ConsistencyBase model=GPT-OSS2026.052
DeepSeek-R1-Distill-Llama-8B2026.071.4
Llama-3.1-8B-Instruct2026.070.6