Math on AMC23 (Score %)
78.3ScoreMPO+
Evaluation Results
| Method | Links | |
|---|---|---|
| MPO+Backbone=olmo-3.1-32b-it-sft2026.04 | 78.3 | |
| Soft+Backbone=olmo-3.1-32b-it-sft2026.04 | 77.5 | |
| Pair+Backbone=olmo-3.1-32b-it-sft2026.04 | 77.2 | |
| Mag+Backbone=olmo-3.1-32b-it-sft2026.04 | 76.6 | |
| MPO+Backbone=olmo-3-7b-it-sft2026.04 | 73.1 | |
| Soft+Backbone=olmo-3-7b-it-sft2026.04 | 72.8 | |
| Pair+Backbone=olmo-3-7b-it-sft2026.04 | 69.5 | |
| RFTBackbone=olmo-3.1-32b-it-sft2026.04 | 68 | |
| Mag+Backbone=olmo-3-7b-it-sft2026.04 | 67.8 | |
| DPOBackbone=olmo-3.1-32b-it-sft2026.04 | 67.7 | |
| DPOBackbone=olmo-3-7b-it-sft2026.04 | 61.6 | |
| TSV-MergeModel Category=Merged Models2026.06 | 55.5 | |
| TAModel Category=Merged Models2026.06 | 55 | |
| RESMERGEModel Category=Merged Models2026.06 | 55 | |
| RFTBackbone=olmo-3-7b-it-sft2026.04 | 54.7 | |
| TIESModel Category=Merged Models2026.06 | 52.75 | |
| RAMModel Category=Merged Models2026.06 | 52.75 | |
| DARE + TIESModel Category=Merged Models2026.06 | 52.5 | |
| ISO-CModel Category=Merged Models2026.06 | 52.5 | |
| BaseBackbone=olmo-3.1-32b-it-sft2026.04 | 52.3 | |
| AMPOModel Category=Base and Expert Models2026.06 | 51.67 | |
| ToolRLModel Category=Base and Expert Models2026.06 | 49.75 | |
| Qwen2.5-7B-InstructModel Category=Base and Expert Models2026.06 | 49.5 | |
| ISO-CTSModel Category=Merged Models2026.06 | 47.5 | |
| BaseBackbone=olmo-3-7b-it-sft2026.04 | 42 | |
| MemoryAgentModel Category=Base and Expert Models2026.06 | 22.25 | |
| Mag+Backbone=gemma-3-4b-sft2026.04 | 16.7 | |
| MPO+Backbone=gemma-3-4b-sft2026.04 | 16.6 | |
| DPOBackbone=gemma-3-4b-sft2026.04 | 16.2 | |
| Pair+Backbone=gemma-3-4b-sft2026.04 | 15.6 | |
| BaseBackbone=gemma-3-4b-sft2026.04 | 15 | |
| RFTBackbone=gemma-3-4b-sft2026.04 | 14.7 | |
| Soft+Backbone=gemma-3-4b-sft2026.04 | 14.5 |