Mathematical Reasoning on MATH (Accuracy + maj1@k)
33.6AccuracyMinerva
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MinervaParams=540B, k=642023.02 | 33.6 | 50.3 | — | |
| MinervaParams=62B, k=2562023.02 | 27.6 | 43.4 | — | |
| MinervaParams=8B, k=2562023.02 | 14.1 | 25.4 | — | |
| LLaMAParams=65B, k=2562023.02 | 10.6 | 20.5 | — | |
| PaLMParams=540B2023.02 | 8.8 | — | — | |
| LLaMAParams=33B, k=2562023.02 | 7.1 | 15.2 | — | |
| PaLMParams=62B2023.02 | 4.4 | — | — | |
| LLaMAParams=13B, k=2562023.02 | 3.9 | 8.8 | — | |
| LLaMAParams=7B, k=2562023.02 | 2.9 | 6.9 | — | |
| PaLMParams=8B2023.02 | 1.5 | — | — | |
| Qwen2.5-Omni-7BModel Category=Perception-centric, Model Scale=7B2026.03 | 0.715 | — | — | |
| Ministral-3-8BModel Category=Text-Experts, Model Scale=8B2026.03 | 0.626 | — | — | |
| Qwen3-8BModel Category=Text-Experts, Model Scale=8B2026.03 | 0.608 | — | — | |
| Dynin-OmniModel Category=Unified, Native speech I/O=true, Block size=16, Diffusion steps=10242026.03 | 0.496 | — | — | |
| HyperCLOVAX-8B-OmniModel Category=Unified, Model Scale=8B, Native speech I/O=true, Result Reproducibility=reproduced2026.03 | 0.434 | — | — | |
| MMaDAModel Category=Unified2026.03 | 0.36 | — | — | |
| Trida-7BModel Category=Text-Experts, Model Scale=7B2026.03 | 0.336 | — | — | |
| LLaDA-8BModel Category=Text-Experts, Model Scale=8B2026.03 | 0.273 | — | — | |
| Llama-3-8BModel Category=Text-Experts, Model Scale=8B2026.03 | 0.205 | — | — | |
| Chameleon-7BModel Category=Unified, Model Scale=7B2026.03 | 0.115 | — | — | |
| FalconSize=7B2023.07 | — | 2.3 | — | |
| FalconSize=40B2023.07 | — | 5.5 | — | |
| LLAMA 1Size=7B2023.07 | — | 2.9 | — | |
| LLAMA 1Size=13B2023.07 | — | 3.9 | — | |
| LLAMA 1Size=33B2023.07 | — | 7.1 | — | |
| LLAMA 1Size=65B2023.07 | — | 10.6 | — | |
| LLAMA 2Size=7B2023.07 | — | 2.5 | — | |
| LLAMA 2Size=13B2023.07 | — | 3.9 | — | |
| LLAMA 2Size=34B2023.07 | — | 6.24 | — | |
| LLAMA 2Size=70B2023.07 | — | 13.5 | — | |
| MPTSize=7B2023.07 | — | 3 | — | |
| MPTSize=30B2023.07 | — | 3.1 | — | |
| MSV64Best-of-K=642026.03 | — | — | 0.7607 | |
| PRM-7BBest-of-K=64, Backbone=Qwen2.5-Math-7B2026.03 | — | — | 0.7411 | |
| PRM-7B + WVBest-of-K=64, Backbone=Qwen2.5-Math-7B, Weighted Voting=true2026.03 | — | — | 0.7433 |