List your company

Accuracy on MMLU Reasoning

91.35AccuracyDMoA
42.83455.429568.02580.6205Nov 17, 2025Dec 21, 2025Jan 24, 2026Feb 27, 2026Apr 2, 2026May 6, 2026Jun 9, 2026

Evaluation results

MethodLinks
DMoABackbone=gpt-oss-120b, Evaluation Paradigm=few-shot2026.0591.35
DR-Rubric-8B (BS-3)Training=SFT+RL, 3K2026.0585.3
ARG-DesignerBackbone=gpt-oss-120b, Evaluation Paradigm=few-shot2026.0585.04
SafeSieveBackbone=gpt-oss-120b2026.0584.65
G-DesignerBackbone=gpt-oss-120b, Evaluation Paradigm=few-shot2026.0584.63
STEERBackbone=gpt-oss-120b2026.0584.4
DALAToken Cons.=3.80 × 10⁵2025.1184.3
SpecReasonBackbone=gpt-oss-120b2026.0584.2
AgentPrune-RToken Cons.=9.60 × 10⁵2025.1184
GPTSwarmBackbone=gpt-oss-120b2026.0583.8
WebExplorer-8BTraining=SFT+RL, 25K2026.0583.8
DR-Rubric-8B (Gemini)Training=SFT+RL, 1K2026.0583.8
LLM-DebateToken Cons.=1.50 × 10⁶2025.1183.7
PHPToken Cons.=2.60 × 10⁶2025.1183.4
LLM-DebateBackbone=gpt-oss-120b2026.0583.28
AutoGenBackbone=gpt-oss-120b2026.0583.25
AFlowBackbone=gpt-oss-120b2026.0583.22
Random GraphBackbone=gpt-oss-120b2026.0583.2
SCBackbone=gpt-oss-120b2026.0582.95
Complete GraphBackbone=gpt-oss-120b2026.0582.75
MoABackbone=gpt-oss-120b2026.0582.65
ChainBackbone=gpt-oss-120b2026.0582.4
VanillaToken Cons.=1.50 × 10⁵2025.1182.1
CoTBackbone=gpt-oss-120b2026.0581.85
Qwen3-8B-SFTTraining=SFT, 1K2026.0581.8
TreeBackbone=gpt-oss-120b2026.0581.55
VanillaBackbone=gpt-oss-120b2026.0580.47
DR-Rubric-8B (GPT-5)Training=SFT+RL, 1K2026.0580.3
DyLANToken Cons.=1.20 × 10⁶2025.1180.2
Relational AggregationCategory=Ours2026.0579.8
StarBackbone=gpt-oss-120b2026.0579.64
Qwen2.5-7B2026.0579.3
GoAMaxCategory=Multi-Agent2026.0579.18
DR-Tulu-SFT-8BTraining=SFT, 16K2026.0579
DR-Tulu-RL-8BTraining=SFT+RL, 25K2026.0579
GoAMeanCategory=Multi-Agent2026.0578.52
Qwen3-8B2026.0578.5
Anchor (Refined Belief)Category=Ours2026.0578.43
Self-MoACategory=Multi-Agent2026.0578.14
Search-R1-7BTraining=RL, 90K2026.0578
SCCategory=Multi-Agent2026.0577.97
GeneralCategory=Single-Agent2026.0577.61
RefineCategory=Multi-Agent2026.0577.4
Anchor (Initial Belief)Category=Ours2026.0577.31
Majority (Refined Belief)Category=Ours2026.0576.22
MoACategory=Multi-Agent2026.0575.71
Majority (Initial Belief)Category=Ours2026.0573.51
DebateCategory=Multi-Agent2026.0572.53
ReConcileCategory=Multi-Agent2026.0569.61
CodeCategory=Single-Agent2026.0568.04
MathCategory=Single-Agent2026.0563.47
LegalCategory=Single-Agent2026.0555.86
FP16Precision=FP162026.0655.26
FinanceCategory=Single-Agent2026.0554.11
BiomedicalCategory=Single-Agent2026.0546.6
LC-QATQuantization=Linear-Constrained Vector Quantization2026.0645.92
PV-TUNING2026.0644.7