Multi-Agent System Performance on Coding
65TS ScoreLLMA-Mem
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LLMA-MemModel=Claude-Sonnet-4.52026.03 | 65 | 44.75 | 54.02 | |
| W.o. MemoryModel=Claude-Sonnet-4.52026.03 | 64.43 | 42.29 | 53.42 | |
| MARBLEModel=Claude-Sonnet-4.52026.03 | 63.05 | 44.71 | 53.99 | |
| A-MemModel=Claude-Sonnet-4.52026.03 | 61.57 | 46.41 | 53.76 | |
| LLMA-MemModel=Qwen3-next-80B2026.03 | 60.65 | 44.86 | 53.12 | |
| LLMA-MemModel=Deepseek-v3.22026.03 | 59.72 | 46.37 | 53.46 | |
| W.o. MemoryModel=Qwen3-next-80B2026.03 | 58.98 | 45.2 | 52.85 | |
| W.o. MemoryModel=Deepseek-v3.22026.03 | 58.69 | 43.47 | 52.26 | |
| A-MemModel=Qwen3-next-80B2026.03 | 57.23 | 48.09 | 52.9 | |
| MARBLEModel=Qwen3-next-80B2026.03 | 57.03 | 46.63 | 52.17 | |
| MARBLEModel=Deepseek-v3.22026.03 | 56.93 | 40.97 | 50.86 | |
| A-MemModel=Deepseek-v3.22026.03 | 49.74 | 37.95 | 46.72 | |
| LLMA-MemModel=Qwen3-32B-Instruct2026.03 | 49.68 | 46.69 | 49.43 | |
| W.o. MemoryModel=Qwen3-32B-Instruct2026.03 | 49.24 | 44.23 | 48.01 | |
| MARBLEModel=Qwen3-32B-Instruct2026.03 | 48.82 | 44.4 | 46.76 | |
| A-MemModel=Qwen3-32B-Instruct2026.03 | 47.05 | 46.79 | 47.66 |