State Abstraction on AMA-Bench SA
52.92Success MetricOPD-Evolver-9B
Evaluation Results
| Method | Links | |
|---|---|---|
| OPD-Evolver-9BBackbone LLM=QWEN3.5-9B2026.06 | 52.92 | |
| QWEN3.5-397B-A17BBackbone LLM=QWEN3.5-397B-A17B2026.06 | 51.41 | |
| STEP-3.5-FLASH (196B)Backbone LLM=STEP-3.5-FLASH (196B)2026.06 | 48.88 | |
| AWMBackbone LLM=QWEN3.5-9B2026.06 | 48 | |
| MemEvolveBackbone LLM=QWEN3.5-9B2026.06 | 47.2 | |
| ReasoningBankBackbone LLM=QWEN3.5-9B2026.06 | 47.04 | |
| MempBackbone LLM=QWEN3.5-9B2026.06 | 46.59 | |
| CheatsheetBackbone LLM=QWEN3.5-9B2026.06 | 46.19 | |
| EvolveRBackbone LLM=QWEN3.5-9B2026.06 | 46.11 | |
| No MemoryBackbone LLM=QWEN3.5-9B2026.06 | 45.63 | |
| ExpeLBackbone LLM=QWEN3.5-9B2026.06 | 44.67 | |
| OPD-Evolver-4BBackbone LLM=QWEN3-4B2026.06 | 34.9 | |
| AWMBackbone LLM=QWEN3-4B2026.06 | 33.49 | |
| MemEvolveBackbone LLM=QWEN3-4B2026.06 | 32.73 | |
| MempBackbone LLM=QWEN3-4B2026.06 | 31.81 | |
| ReasoningBankBackbone LLM=QWEN3-4B2026.06 | 30.97 | |
| CheatsheetBackbone LLM=QWEN3-4B2026.06 | 30.37 | |
| EvolveRBackbone LLM=QWEN3-4B2026.06 | 30.01 | |
| No MemoryBackbone LLM=QWEN3-4B2026.06 | 29.73 | |
| ExpeLBackbone LLM=QWEN3-4B2026.06 | 28.49 |