Code Generation on LCB v5 (Accuracy)
61.5AccuracyQwen3-30B-A3B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-30B-A3BBase Model=Qwen3-30B-A3B2026.05 | 61.5 | |
| ZEDASFTBase Model=Qwen3-30B-A3B2026.05 | 58.2 | |
| ZEDABase Model=Qwen3-30B-A3B2026.05 | 58.2 | |
| Dynamic SkippingBase Model=Qwen3-30B-A3B2026.05 | 57.3 | |
| NETSFT→OPDBase Model=Qwen3-30B-A3B2026.05 | 57 | |
| NETSFTBase Model=Qwen3-30B-A3B2026.05 | 56.5 | |
| ZEDABase Model=GLM-4.7-Flash2026.05 | 51.6 | |
| NETSFTBase Model=GLM-4.7-Flash2026.05 | 50.6 | |
| ZEDASFTBase Model=GLM-4.7-Flash2026.05 | 49.9 | |
| NETSFT→OPDBase Model=GLM-4.7-Flash2026.05 | 49.5 | |
| GLM-4.7-FlashBase Model=GLM-4.7-Flash2026.05 | 48 | |
| Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Base+Tool2026.04 | 39.76 | |
| ParaManagerOrchestration Strategy=Unified2026.04 | 39.45 | |
| ParaManager-SerialOrchestration Strategy=Serial2026.04 | 39.43 | |
| Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Self Refine2026.04 | 39.42 | |
| Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Majority Vote2026.04 | 39.28 | |
| ParaManager-MonoOrchestration Strategy=Mono2026.04 | 39.2 | |
| RouterOrchestration Strategy=Static Workflow2026.04 | 39.15 | |
| Qwen3-30B A3B-Instruct-2507Orchestration Strategy=Base2026.04 | 39.12 | |
| PuppeteerOrchestration Strategy=Serial Orchestration2026.04 | 39.11 | |
| Qwen3-4B Instruct-2507Orchestration Strategy=Self Refine2026.04 | 38.36 | |
| ToolOrchestraOrchestration Strategy=Serial Orchestration2026.04 | 37.64 | |
| Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Self Refine2026.04 | 37.6 | |
| GPT-OSS-20BOrchestration Strategy=Self Refine2026.04 | 37.2 | |
| Qwen3-4B Instruct-2507Orchestration Strategy=Base+Tool2026.04 | 37.15 | |
| Qwen3-4B Instruct-2507Orchestration Strategy=Base2026.04 | 36.6 | |
| Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Base+Tool2026.04 | 36.2 | |
| AdaMoEBase Model=Qwen3-30B-A3B2026.05 | 36.1 | |
| Qwen3-4B Instruct-2507Orchestration Strategy=Majority Vote2026.04 | 35.83 | |
| Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Majority Vote2026.04 | 35.2 | |
| Qwen3-Coder 30B-A3B-InstructOrchestration Strategy=Base2026.04 | 34.76 | |
| GPT-OSS-20BOrchestration Strategy=Base+Tool2026.04 | 34.58 | |
| GPT-OSS-20BOrchestration Strategy=Majority Vote2026.04 | 34.5 | |
| GPT-OSS-20BOrchestration Strategy=Base2026.04 | 33.53 | |
| EvoFlowOrchestration Strategy=Static Workflow2026.04 | 33.06 | |
| Dynamic SkippingBase Model=GLM-4.7-Flash2026.05 | 32.3 | |
| Meta Agent SearchOrchestration Strategy=Static Workflow2026.04 | 32.08 | |
| STEERModel Size=14B2025.10 | 31.8 | |
| ParaManager-SFTOrchestration Strategy=SFT2026.04 | 30.8 | |
| GRPOModel Size=14B2025.10 | 29.3 | |
| STEERModel Size=7B2025.10 | 29.2 | |
| GRPOModel Size=7B2025.10 | 28.5 | |
| AdaMoEBase Model=GLM-4.7-Flash2026.05 | 26.4 | |
| STEERModel Size=3B2025.10 | 24.9 | |
| GRPOModel Size=3B2025.10 | 24.4 |