List your company

Code Generation on HumanEval (Pass@k, AUC, and Token Usage)

94.1Pass@1ExPairT-LLM
45.53258.14170.7583.359May 26, 2025Jul 14, 2025Sep 2, 2025Oct 21, 2025Dec 10, 2025Jan 28, 2026Mar 19, 2026

Evaluation results

MethodLinks
ExPairT-LLMModel=CodeGen2025.1194.1—————
Vanilla (Ref)LLM=Claude-3.5-Sonnet, Multi-agent=No, Dynamic Routing=No2026.0393.7—————
MaASDyn.=true, Dis.=false2025.1192.85—————
AMRO-SLLM=LLM Pool*, Multi-agent=Yes, Dynamic Routing=Yes2026.0392.2—————
OneFlowMethod category=Single-LLM implementation, Executor LLM=GPT-4o-mini, Execution protocol=single-agent execution2026.0192.1—————
ExPairT-LLMModel=StarCoder2025.1191.7—————
OneFlowMethod category=Automatically designed multi-agent frameworks, Executor LLM=GPT-4o-mini, Execution protocol=multi-agent2026.0191.6—————
ExPairT-LLMModel=DeepSeek-Coder2025.1191.6—————
BiRouterDyn.=true, Dis.=true2025.1191.46—————
ExPairT-LLMModel=Codex2025.1191.3—————
AFlowMethod category=Single-LLM implementation, Executor LLM=GPT-4o-mini, Execution protocol=single-agent execution2026.0191.1—————
MasRouterLLM=LLM Pool*, Multi-agent=Yes, Dynamic Routing=Yes2026.0390.62—————
DyLANDyn.=true, Dis.=false2025.1190.42—————
CoTMethod category=Manual baselines, Executor LLM=GPT-4o-mini, Execution protocol=standard2026.0190.3—————
Vanilla (Ref)LLM=GPT-4o, Multi-agent=No, Dynamic Routing=No2026.0390.2—————
AFlowMethod category=Automatically designed multi-agent frameworks, Executor LLM=GPT-4o-mini, Execution protocol=multi-agent2026.0190.1—————
AFlowLLM=GPT-4o-mini, Multi-agent=Yes, Dynamic Routing=No2026.0390.06—————
CoT SCMethod category=Manual baselines, Executor LLM=GPT-4o-mini, Execution protocol=standard, shots=5-shot2026.0189.8—————
GPTSwarmDyn.=false, Dis.=false2025.1189.32—————
AgentVerseDyn.=true, Dis.=false2025.1189.29—————
IOMethod category=Manual baselines, Executor LLM=GPT-4o-mini, Execution protocol=standard2026.0189.1—————
MultiPersonaMethod category=Manual baselines, Executor LLM=GPT-4o-mini, Execution protocol=standard2026.0189.1—————
ExPairT-LLMModel=CodeLlama2025.1189.1—————
LLM-BlenderDyn.=false, Dis.=false2025.1188.8—————
LLM-DebateDyn.=false, Dis.=false2025.1188.68—————
SC (CoTx5)2025.1188.6—————
CoT2025.1188.13—————
BaseBackbone=Qwen3-8B2025.0988—————
GRPOBackbone=Qwen3-8B2025.0987.8—————
RouterDCLLM=LLM Pool*, Multi-agent=No, Dynamic Routing=Yes2026.0387.75—————
G-DesignerDyn.=false, Dis.=false2025.1187.5—————
AoT (Atom)LLM=GPT-4o-mini, Multi-agent=No, Dynamic Routing=No2026.0387.5—————
ComplexCoT2025.1187.49—————
Single-agent2025.1187.08—————
IPOBackbone=Qwen3-8B2025.0986.8—————
CoTLLM=GPT-4o-mini, Multi-agent=No, Dynamic Routing=No2026.0386.69—————
VanillaLLM=Claude-3.5-haiku, Multi-agent=No, Dynamic Routing=No2026.0386.4—————
GPTSwarmLLM=GPT-4o-mini, Multi-agent=Yes, Dynamic Routing=No2026.0386.28—————
GoT (Graph)LLM=GPT-4o-mini, Multi-agent=No, Dynamic Routing=No2026.0385.8—————
VanillaLLM=GPT-4o-mini, Multi-agent=No, Dynamic Routing=No2026.0385.71—————
AFlowLLM=Gemini-1.5-flash, Multi-agent=Yes, Dynamic Routing=No2026.0385.69—————
ToT (Tree)LLM=GPT-4o-mini, Multi-agent=No, Dynamic Routing=No2026.0385.2—————
MacNetDyn.=false, Dis.=false2025.1184.57—————
LLM-DebateLLM=GPT-4o-mini, Multi-agent=Yes, Dynamic Routing=No2026.0384.38—————
RouteLLMLLM=LLM Pool*, Multi-agent=No, Dynamic Routing=Yes2026.0383.85—————
VanillaLLM=Gemini-1.5-flash, Multi-agent=No, Dynamic Routing=No2026.0382.61—————
IPOBackbone=DeepSeek-R1-Distill-Qwen-7B2025.0982.6—————
SafeKeyBackbone=DeepSeek-R1-Distill-Llama-8B2025.0981.9—————
BaseBackbone=DeepSeek-R1-Distill-Qwen-7B2025.0981.7—————
SafeKeyBackbone=DeepSeek-R1-Distill-Qwen-7B2025.0981—————
GRPOBackbone=DeepSeek-R1-Distill-Qwen-7B2025.0981—————
CODETModel=DeepSeek-Coder2025.1180.9—————
B^4Model=DeepSeek-Coder2025.1180.8—————
B^4Model=Codex2025.1180.5—————
VanillaLLM=Llama-3.1-70b, Multi-agent=No, Dynamic Routing=No2026.0380.5—————
STARBackbone=DeepSeek-R1-Distill-Qwen-7B2025.0980—————
GRPOBackbone=DeepSeek-R1-Distill-Llama-8B2025.0979.9—————
BaseBackbone=DeepSeek-R1-Distill-Llama-8B2025.0979.5—————
IPOBackbone=DeepSeek-R1-Distill-Llama-8B2025.0979.5—————
RealSafeBackbone=DeepSeek-R1-Distill-Qwen-7B2025.0979.2—————
RealSafeBackbone=DeepSeek-R1-Distill-Llama-8B2025.0977.8—————
SafeChainBackbone=DeepSeek-R1-Distill-Llama-8B2025.0977.4—————
STARBackbone=DeepSeek-R1-Distill-Llama-8B2025.0977.1—————
ReasoningBackbone=Qwen2.5-7B family2026.0176.63—————
SafeChainBackbone=DeepSeek-R1-Distill-Qwen-7B2025.0976.5—————
CODETModel=Codex2025.1174.1—————
Claude-Haiku-4.5Category=Candidate Models2026.0373.7—————
B^4Model=StarCoder2025.1170.5—————
FineRouterCategory=Routers2026.0368.4—————
TI-DPOModel=LLaMA-3.2-3B2025.0568—————
B^4Model=CodeLlama2025.1166.6—————
CODETModel=StarCoder2025.1165.4—————
OriginalModel=DeepSeek-Coder2025.1165.4—————
B^4Model=CodeGen2025.1164—————
TPOModel=LLaMA-3.2-3B2025.0564—————
GRPOModel=LLaMA-3.2-3B2025.0563.8—————
CPOModel=LLaMA-3.2-3B2025.0563.5—————
Claude-Sonnet-4.5Category=Candidate Models2026.0363.2—————
MLPCategory=Routers2026.0363.2—————
KTOModel=LLaMA-3.2-3B2025.0563—————
CODETModel=CodeLlama2025.1162.5—————
DPOModel=LLaMA-3.2-3B2025.0562—————
TDPOModel=LLaMA-3.2-3B2025.0562—————
ReasonAnyBackbone=Qwen2.5-7B family2026.0161.71—————
SIMPOModel=LLaMA-3.2-3B2025.0561.5—————
SFTModel=LLaMA-3.2-3B2025.0561—————
IPOModel=LLaMA-3.2-3B2025.0559—————
Mistral-LargeCategory=Candidate Models2026.0357.9—————
Dele-SimKOBackbone=Qwen3-8B2026.0157.180.583.573.25.8—
Dele-GRPOBackbone=Qwen3-8B2026.0156.479.882.571.25.9—
LCoT-SimKOBackbone=Qwen3-8B2026.0156.279.28271.86—
TGR-LatentBackbone=Qwen3-8B2026.0156.181.184.273.54.6—
TGR-TokenBackbone=Qwen3-8B2026.0155.378.581.870.84.4—
CODETModel=CodeGen2025.1155.1—————
LCoT-GRPOBackbone=Qwen3-8B2026.015576.879.569.56—
PSamplingBackbone=Qwen3-8B2026.0152.874.577.867.23.6—
Qwen3-235B-A22BCategory=Candidate Models2026.0352.6—————
DAREBackbone=Qwen2.5-7B family2026.0149.4—————
IOA (Ours)Model=Qwen2.5-7B2026.0247.67—————
DeepSeek-R1Category=Candidate Models2026.0347.4—————