Math on OlympiadBench (Score, TPF)
77.59ScoreQwen3-30B-A3B-Inst-2507
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-30B-A3B-Inst-25072026.02 | 77.59 | 100 | |
| LLaDA2.1-flashInference Mode=Q Mode2026.02 | 76.59 | 381 | |
| Ling-flash-2.02026.02 | 76.19 | 100 | |
| LLaDA2.1-flashInference Mode=S Mode2026.02 | 75.85 | 646 | |
| LLaDA2.0-flash2026.02 | 74.07 | 370 | |
| Ling-mini-2.02026.02 | 72.3 | — | |
| LLaDA2.0-mini2026.02 | 67.7 | 2.63 | |
| Pair+Backbone=olmo-3.1-32b-it-sft2026.04 | 67 | — | |
| LLaDA2.1-minimode=Q Mode2026.02 | 66.67 | 3.99 | |
| MPO+Backbone=olmo-3.1-32b-it-sft2026.04 | 64.7 | — | |
| LLaDA2.1-minimode=S Mode2026.02 | 64.3 | 7.08 | |
| Soft+Backbone=olmo-3.1-32b-it-sft2026.04 | 64.1 | — | |
| Mag+Backbone=olmo-3.1-32b-it-sft2026.04 | 61.3 | — | |
| DPOBackbone=olmo-3.1-32b-it-sft2026.04 | 59.9 | — | |
| MPO+Backbone=olmo-3-7b-it-sft2026.04 | 57.3 | — | |
| Pair+Backbone=olmo-3-7b-it-sft2026.04 | 57.3 | — | |
| Soft+Backbone=olmo-3-7b-it-sft2026.04 | 57.3 | — | |
| Mag+Backbone=olmo-3-7b-it-sft2026.04 | 56 | — | |
| Qwen3-8Bno think=true2026.02 | 55.33 | — | |
| RFTBackbone=olmo-3.1-32b-it-sft2026.04 | 52.9 | — | |
| DPOBackbone=olmo-3-7b-it-sft2026.04 | 51.4 | — | |
| RFTBackbone=olmo-3-7b-it-sft2026.04 | 43.7 | — | |
| BaseBackbone=olmo-3.1-32b-it-sft2026.04 | 43.7 | — | |
| BaseBackbone=olmo-3-7b-it-sft2026.04 | 34.5 | — | |
| Pair+Backbone=gemma-3-4b-sft2026.04 | 13.5 | — | |
| MPO+Backbone=gemma-3-4b-sft2026.04 | 13.2 | — | |
| Mag+Backbone=gemma-3-4b-sft2026.04 | 11.9 | — | |
| RFTBackbone=gemma-3-4b-sft2026.04 | 11 | — | |
| DPOBackbone=gemma-3-4b-sft2026.04 | 10.8 | — | |
| Soft+Backbone=gemma-3-4b-sft2026.04 | 10.8 | — | |
| BaseBackbone=gemma-3-4b-sft2026.04 | 9.6 | — |