Code Generation on LCB v6
57.1AccuracyQwen3-30B-A3B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-30B-A3BBase Model=Qwen3-30B-A3B2026.05 | 57.1 | |
| ZEDABase Model=Qwen3-30B-A3B2026.05 | 53.2 | |
| NETSFT→OPDBase Model=Qwen3-30B-A3B2026.05 | 52.9 | |
| ZEDASFTBase Model=Qwen3-30B-A3B2026.05 | 52.8 | |
| Dynamic SkippingBase Model=Qwen3-30B-A3B2026.05 | 51.9 | |
| NETSFTBase Model=Qwen3-30B-A3B2026.05 | 50.9 | |
| SDPO2026.01 | 48.8 | |
| Qwen2.5-32B-Instruct + Bootcamp-SFT-RLModel=Qwen2.5-32B-Instruct, Training Stage=Bootcamp-SFT-RL2025.08 | 47.4 | |
| ZEDABase Model=GLM-4.7-Flash2026.05 | 45.6 | |
| ZEDASFTBase Model=GLM-4.7-Flash2026.05 | 45 | |
| NETSFTBase Model=GLM-4.7-Flash2026.05 | 44.7 | |
| DS-R1-Distilled-Qwen-32B + Bootcamp-RLModel=DS-R1-Distilled-Qwen-32B, Training Stage=Bootcamp-RL2025.08 | 44.6 | |
| GLM-4.7-FlashBase Model=GLM-4.7-Flash2026.05 | 44.4 | |
| Qwen2.5-32B-Instruct + Bootcamp-SFTModel=Qwen2.5-32B-Instruct, Training Stage=Bootcamp-SFT2025.08 | 44 | |
| NETSFT→OPDBase Model=GLM-4.7-Flash2026.05 | 43.8 | |
| Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Self Refine2026.04 | 43.25 | |
| Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Base+Tool2026.04 | 42.8 | |
| SFT on self-teacher2026.01 | 42.7 | |
| ParaManager-MonoOrchestration Strategy=Mono2026.04 | 42.5 | |
| Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Majority Vote2026.04 | 42.45 | |
| Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Base2026.04 | 42.1 | |
| ParaManagerOrchestration Strategy=Unified2026.04 | 41.51 | |
| ParaManager-SerialOrchestration Strategy=Serial2026.04 | 41.38 | |
| Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Base+Tool2026.04 | 41.25 | |
| RouterOrchestration Strategy=Static Workflow2026.04 | 41.25 | |
| GRPO2026.01 | 41.2 | |
| PuppeteerOrchestration Strategy=Serial Orchestration2026.04 | 41.06 | |
| GPT-OSS-20BOrchestration Strategy=Base+Tool2026.04 | 40.87 | |
| GPT-OSS-20BOrchestration Strategy=Self Refine2026.04 | 40.2 | |
| Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Majority Vote2026.04 | 40.2 | |
| Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Self Refine2026.04 | 40.2 | |
| Qwen3-4B Instruct-2507Orchestration Strategy=Self Refine2026.04 | 40.12 | |
| Qwen3-4B Instruct-2507Orchestration Strategy=Base2026.04 | 39.95 | |
| Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Base2026.04 | 39.68 | |
| Qwen3-4B Instruct-2507Orchestration Strategy=Majority Vote2026.04 | 39.54 | |
| Qwen3-4B Instruct-2507Orchestration Strategy=Base+Tool2026.04 | 39.52 | |
| ToolOrchestraOrchestration Strategy=Serial Orchestration2026.04 | 38.27 | |
| GPT-OSS-20BOrchestration Strategy=Base2026.04 | 37.43 | |
| GPT-OSS-20BOrchestration Strategy=Majority Vote2026.04 | 37.2 | |
| EvoFlowOrchestration Strategy=Static Workflow2026.04 | 37.12 | |
| Meta Agent SearchOrchestration Strategy=Static Workflow2026.04 | 35.16 | |
| AdaMoEBase Model=Qwen3-30B-A3B2026.05 | 34.3 | |
| DS-R1-Distilled-Qwen-32BModel=DS-R1-Distilled-Qwen-32B2025.08 | 33.7 | |
| ParaManager-SFTOrchestration Strategy=SFT2026.04 | 33.58 | |
| Dynamic SkippingBase Model=GLM-4.7-Flash2026.05 | 32.4 | |
| Qwen2.5-32B-Instruct + Bootcamp-RLModel=Qwen2.5-32B-Instruct, Training Stage=Bootcamp-RL2025.08 | 29.7 | |
| AdaMoEBase Model=GLM-4.7-Flash2026.05 | 28.6 | |
| Qwen2.5-32B-InstructModel=Qwen2.5-32B-Instruct2025.08 | 28 | |
| Base2026.01 | 27.9 | |
| SDAR-8BScale=8B, Tokens=55B, FLOPs=2640, Decoding Strategy=greedy static decoding, Block Size=42026.06 | 14.5 | |
| SDAR-4BScale=4B, Tokens=55B, FLOPs=1320, Decoding Strategy=greedy static decoding, Block Size=42026.06 | 12.6 | |
| OPDLM-4BScale=4B, Tokens=0.076B, FLOPs=2.4, Decoding Strategy=greedy static decoding, Block Size=42026.06 | 10.4 | |
| Fast-dLLM-v2-7BScale=8B, Tokens=1B, FLOPs=42, Decoding Strategy=greedy static decoding, Block Size=42026.06 | 9.7 | |
| OPDLM-8BScale=8B, Tokens=0.066B, FLOPs=4.2, Decoding Strategy=greedy static decoding, Block Size=42026.06 | 9.7 |