Mathematical Reasoning on OMNI-MATH (Overall Accuracy)
39.55Overall AccuracyTEPO
Evaluation Results
| Method | Links | |
|---|---|---|
| TEPOBackbone=DeepSeek-R1-Distill-Qwen-7B2026.04 | 39.55 | |
| GRPO/DAPOBackbone=DeepSeek-R1-Distill-Qwen-7B2026.04 | 38.42 | |
| TEPOBackbone=Qwen3-14B, Method Variant=TEPO2026.04 | 35.11 | |
| DeepSeek-R1-Distill-Qwen-7BBackbone=DeepSeek-R1-Distill-Qwen-7B2026.04 | 33.86 | |
| Qwen3-14B w. GSPOBackbone=Qwen3-14B, Method Variant=GSPO2026.04 | 33.54 | |
| TEPOBackbone=DeepSeek-R1-Distill-Llama-8B2026.04 | 33.11 | |
| Qwen3-14B w. Entropy-based TermBackbone=Qwen3-14B, Method Variant=Entropy-based Term2026.04 | 33.04 | |
| Qwen3-14B w. GRPO/DAPOBackbone=Qwen3-14B, Method Variant=GRPO/DAPO2026.04 | 33.01 | |
| GRPO/DAPOBackbone=DeepSeek-R1-Distill-Llama-8B2026.04 | 32.97 | |
| Qwen3-14B w. CLIP-CovBackbone=Qwen3-14B, Method Variant=CLIP-Cov2026.04 | 32.94 | |
| Qwen3-14B w. KL-CovBackbone=Qwen3-14B, Method Variant=KL-Cov2026.04 | 32.87 | |
| Qwen3-14B w. GPGBackbone=Qwen3-14B, Method Variant=GPG2026.04 | 32.22 | |
| Qwen3-14BBackbone=Qwen3-14B, Method Variant=Base2026.04 | 30.27 | |
| TEPOBackbone=Qwen2.5-7B, Method Variant=TEPO2026.04 | 27.17 | |
| Qwen2.5-7B w. GPGBackbone=Qwen2.5-7B, Method Variant=GPG2026.04 | 26.28 | |
| Qwen2.5-7B w. Entropy-based TermBackbone=Qwen2.5-7B, Method Variant=Entropy-based Term2026.04 | 26.14 | |
| Qwen2.5-7B w. CLIP-CovBackbone=Qwen2.5-7B, Method Variant=CLIP-Cov2026.04 | 26.11 | |
| Qwen2.5-7B w. GSPOBackbone=Qwen2.5-7B, Method Variant=GSPO2026.04 | 25.72 | |
| Qwen2.5-7B w. KL-CovBackbone=Qwen2.5-7B, Method Variant=KL-Cov2026.04 | 25.64 | |
| DeepSeek-R1-Distill-Llama-8BBackbone=DeepSeek-R1-Distill-Llama-8B2026.04 | 25.5 | |
| Qwen2.5-7B w. GRPO/DAPOBackbone=Qwen2.5-7B, Method Variant=GRPO/DAPO2026.04 | 25 | |
| Qwen2.5-7BBackbone=Qwen2.5-7B, Method Variant=Base2026.04 | 13.73 | |
| TEPOBackbone=Mistral-7B-Instruct-v0.22026.04 | 4.59 | |
| GRPO/DAPOBackbone=Mistral-7B-Instruct-v0.22026.04 | 4.27 | |
| Mistral-7B-Instruct-v0.2Backbone=Mistral-7B-Instruct-v0.22026.04 | 3.98 |