List your company
ResearchFormerly WizWand

Knowledge Reasoning on MMLU (Accuracy)

92.16MMLU Knowledge Reasoning AccuracySIGMA
29.91646.075562.23578.3945Oct 21, 2025Nov 25, 2025Dec 30, 2025Feb 3, 2026Mar 10, 2026Apr 14, 2026May 19, 2026
Updated 5mo ago

Evaluation results

MethodLinks
SIGMAMulti-agent collaboration (Mul.)=full, Inter-agent relations (Rel.)=full, Conflicting signals (Conf.)=full, Backbone=DeepSeek-V3.22026.0592.16
GoAMulti-agent collaboration (Mul.)=full, Inter-agent relations (Rel.)=full, Conflicting signals (Conf.)=partial, Backbone=DeepSeek-V3.22026.0591.53
G-DesignerMulti-agent collaboration (Mul.)=full, Inter-agent relations (Rel.)=full, Conflicting signals (Conf.)=no, Backbone=DeepSeek-V3.22026.0590.85
GPTSwarmMulti-agent collaboration (Mul.)=full, Inter-agent relations (Rel.)=full, Conflicting signals (Conf.)=no, Backbone=DeepSeek-V3.22026.0590.03
ST-EVOBackbone=gpt-oss-120b2026.0289.85
MoAMulti-agent collaboration (Mul.)=full, Inter-agent relations (Rel.)=no, Conflicting signals (Conf.)=no, Backbone=DeepSeek-V3.22026.0589.77
ComplexCoTMulti-agent collaboration (Mul.)=no, Inter-agent relations (Rel.)=no, Conflicting signals (Conf.)=no, Backbone=DeepSeek-V3.22026.0589.41
CoTMulti-agent collaboration (Mul.)=no, Inter-agent relations (Rel.)=no, Conflicting signals (Conf.)=no, Backbone=DeepSeek-V3.22026.0588.74
VanillaMulti-agent collaboration (Mul.)=no, Inter-agent relations (Rel.)=no, Conflicting signals (Conf.)=no, Backbone=DeepSeek-V3.22026.0587.46
Qwen3Params=32B2026.0487.2
I-DLMParams=32B, Decoding=ISD (N=4, sampling)2026.0486.8
ARG-DesignerBackbone=gpt-oss-120b2026.0285.04
SafeSieveBackbone=gpt-oss-120b2026.0284.65
G-DesignerBackbone=gpt-oss-120b2026.0284.63
STEERBackbone=gpt-oss-120b2026.0284.4
SpecReasonBackbone=gpt-oss-120b2026.0284.2
LangGraphBackbone=gpt-oss-120b2026.0284.1
GPTSwarmBackbone=gpt-oss-120b2026.0283.8
Qwen3Params=8B2026.0483.5
LLM-DebateBackbone=gpt-oss-120b2026.0283.28
AutoGenBackbone=gpt-oss-120b2026.0283.25
AFlowBackbone=gpt-oss-120b2026.0283.22
Random GraphBackbone=gpt-oss-120b2026.0283.2
SCBackbone=gpt-oss-120b2026.0282.95
SDARParams=30B2026.0482.8
Complete GraphBackbone=gpt-oss-120b2026.0282.75
ChainBackbone=gpt-oss-120b2026.0282.4
I-DLMParams=8B, Decoding=ISD (N=4, sampling)2026.0482.4
CoTBackbone=gpt-oss-120b2026.0281.85
TreeBackbone=gpt-oss-120b2026.0281.55
VanillaBackbone=gpt-oss-120b2026.0280.47
StarBackbone=gpt-oss-120b2026.0279.64
SDARParams=8B2026.0478.6
POESOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0477.9
LLaDA-2.1-miniParams=16B2026.0474.5
AnchorOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0471
IPOMPOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0470.1
SESSOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0469.8
RandomOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0468.8
PredictionOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0468.7
POESOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0466.6
POESOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0466.4
Qwen2.5-VL 3BText Source=Pure-Text, Zero-shot setup=true2025.1061.91
POESOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0459.5
PredictionOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0458.5
POESOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0457.1
PredictionOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0456.7
POESOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0456.5
SESSOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0455.8
RandomOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0455.5
RandomOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0455.2
IPOMPOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0454.9
AnchorOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0454.8
AnchorOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0454.6
IPOMPOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0453.9
SESSOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0453.7
SEETOKText Source=Visual-Text, Zero-shot setup=true2025.1052.52
RandomOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0449.9
PredictionOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0449.1
AnchorOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0448.8
IPOMPOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0448.5
SESSOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0448.1
AnchorOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0448.1
SESSOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0447
PredictionOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0446.5
IPOMPOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0445.9
IPOMPOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0445.5
RandomOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0444.6
AnchorOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0444.2
RandomOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0444
PredictionOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0443.7
SESSOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0443.5
Qwen2.5-VL 3BText Source=Visual-Text, Zero-shot setup=true2025.1032.31