List your company
ResearchFormerly WizWand

General Knowledge Reasoning on MMLU Pro (Accuracy)

82.93AccuracyQwen3-Next-80B-A3B-Instruct
31.938845.176958.41571.6531May 26, 2025Jul 23, 2025Sep 19, 2025Nov 16, 2025Jan 13, 2026Mar 12, 2026May 9, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Qwen3-Next-80B-A3B-Instruct2026.0382.93
ParaManager-SFTOrchestration Strategy=SFT2026.0481.43
ParaManagerOrchestration Strategy=Unified2026.0481.43
PuppeteerOrchestration Strategy=Serial Orchestration2026.0480.54
Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Self Refine2026.0480.36
GPT-OSS-20BOrchestration Strategy=Base+Tool2026.0480
GPT-OSS-20BOrchestration Strategy=Majority Vote2026.0480
GPT-OSS-20BOrchestration Strategy=Self Refine2026.0480
Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Majority Vote2026.0480
RouterOrchestration Strategy=Static Workflow2026.0480
ParaManager-SerialOrchestration Strategy=Serial2026.0480
Qwen3-Omni-A3B-Instruct2026.0379.89
EvoFlowOrchestration Strategy=Static Workflow2026.0479.82
GPT-OSS-20BOrchestration Strategy=Base2026.0479.64
ParaManager-MonoOrchestration Strategy=Mono2026.0479.64
Meta Agent SearchOrchestration Strategy=Static Workflow2026.0479.46
ToolOrchestraOrchestration Strategy=Serial Orchestration2026.0479.46
Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Majority Vote2026.0478.57
Nemotron 3 Nano 30B-A3BOpen-Source=✓, Size=30B-A3B2026.0478.3
Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Base+Tool2026.0477.86
Nemotron 3 Nano OmniOpen-Source=✓, Size=30B-A3B2026.0477.3
Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Base2026.0477.14
LongCat-Next2026.0377.02
Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Base2026.0476.96
Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Self Refine2026.0476.79
Qwen3-4B Instruct-2507Orchestration Strategy=Self Refine2026.0475.71
Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Base+Tool2026.0475.71
Qwen3-4B Instruct-2507Orchestration Strategy=Base2026.0474.82
Qwen3-4B Instruct-2507Orchestration Strategy=Base+Tool2026.0474.11
Qwen3-4B Instruct-2507Orchestration Strategy=Majority Vote2026.0472.86
Kimi-Linear-48B-A3B2026.0367.22
HeRLModel=Qwen3-4B-Instruct-2507, Training Method=HeRL2026.0362.1
Qwen3-OmniOpen-Source=✓, Size=30B-A3B2026.0461.6
Qwen3-4B-Instruct-2507Model=Qwen3-4B-Instruct-2507, Training Method=Instruct2026.0361.5
INTUITORBackbone=Qwen3-14B, Evaluation Protocol=standard evaluation protocol2025.0561.3
Qwen3-14BBackbone=Qwen3-14B, Evaluation Protocol=standard evaluation protocol2025.0559.7
GRPOBackbone=Qwen3-8B-Base2026.0559.05
INTUITORBackbone=Qwen2.5-14B, Evaluation Protocol=standard evaluation protocol2025.0558.3
SPIRALBackbone=Qwen3-8B-Base2026.0558.28
STRATAGEMModel=STRATAGEM (Ours)2026.0457.83
GRPOBackbone=Qwen2.5-14B, Evaluation Protocol=standard evaluation protocol2025.0557.8
MARSBackbone=Qwen3-8B-Base2026.0557.8
DEPTBackbone=Qwen3-8B-Base2026.0557.6
Qwen2.5-7B-InstructModel=Qwen2.5-7B-Instruct, Training Method=Instruct2026.0357.3
Qwen2.5-14BBackbone=Qwen2.5-14B, Evaluation Protocol=standard evaluation protocol2025.0556.5
DEPTBackbone=Qwen3-4B-Base2026.0556.45
HeRLModel=Qwen2.5-7B-Instruct, Training Method=HeRL2026.0355.5
SPIRALModel=SPIRAL2026.0453.93
MARSBackbone=Qwen3-4B-Base2026.0553.34
SPAGBackbone=Qwen3-4B-Base2026.0552.71
SPAGBackbone=Qwen3-8B-Base2026.0551.46
INTUITORBackbone=Qwen2.5-7B, Evaluation Protocol=standard evaluation protocol2025.0551.4
SPIRALBackbone=Qwen3-4B-Base2026.0551.32
GRPOBackbone=Qwen2.5-7B, Evaluation Protocol=standard evaluation protocol2025.0551.1
PSFTTraining Protocol=PSFT2025.0850.55
Qwen2.5-7BBackbone=Qwen2.5-7B, Evaluation Protocol=standard evaluation protocol2025.0549.7
GRPOBackbone=Qwen3-4B-Base2026.0549.14
BaseTraining Protocol=Base2025.0847.9
Qwen3-4B-BaseModel=Qwen3-4B-Base2026.0447.2
VANILLABackbone=Qwen3-8B-Base2026.0546.97
SFTTraining Protocol=SFT2025.0840.67
VANILLABackbone=Qwen3-4B-Base2026.0539.36
HeRLModel=Llama-3.2-3B-Instruct, Training Method=HeRL2026.0334.8
Llama-3.2-3B-InstructModel=Llama-3.2-3B-Instruct, Training Method=Instruct2026.0333.9