Mathematical Reasoning on OlympBench
74.04Pass@1Qwen3-235B-A22B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-235B-A22BModel Category=Large Language Models, Thinking Mode=Thinking2025.12 | 74.04 | |
| FIGRModel Category=Large Vision-Language Models2025.12 | 72.4 | |
| Qwen3-32BModel Category=Large Language Models, Thinking Mode=Thinking2025.12 | 71.96 | |
| Text-only RLModel Category=Large Vision-Language Models, Training=Reinforcement Learning (GRPO)2025.12 | 70.33 | |
| GLM-4.5VModel Category=Large Vision-Language Models, Parameters=108B2025.12 | 70.03 | |
| Qwen3-VL-32B-InstructModel Category=Large Vision-Language Models, Parameters=32B2025.12 | 69.73 | |
| Qwen3-VL-8B-InstructModel Category=Large Vision-Language Models, Parameters=8B2025.12 | 67.21 | |
| Qwen3-32BModel Category=Large Language Models, Thinking Mode=Non-Thinking2025.12 | 52.08 | |
| RELEXBackbone=Qwen3-8B-Base, Cost=20%2026.05 | 49.6 | |
| RLVRBackbone=Qwen3-8B-Base, Cost=100%2026.05 | 49.5 | |
| Weight Extrap.Backbone=Qwen3-8B-Base, Cost=20%2026.05 | 48.5 | |
| RELEXBackbone=Qwen3-4B-Base, Cost=15%2026.05 | 47.4 | |
| ExPOBackbone=Qwen3-8B-Base, Cost=20%2026.05 | 47.1 | |
| AlphaRLBackbone=Qwen3-8B-Base, Cost=20%2026.05 | 46.9 | |
| RLVRBackbone=Qwen3-4B-Base, Cost=100%2026.05 | 46.6 | |
| Logits Extrap.Backbone=Qwen3-8B-Base, Cost=20%2026.05 | 46 | |
| ExPOBackbone=Qwen3-4B-Base, Cost=15%2026.05 | 44.1 | |
| Weight Extrap.Backbone=Qwen3-4B-Base, Cost=15%2026.05 | 42.9 | |
| AlphaRLBackbone=Qwen3-4B-Base, Cost=15%2026.05 | 42 | |
| Logits Extrap.Backbone=Qwen3-4B-Base, Cost=15%2026.05 | 40.1 | |
| BaseBackbone=Qwen3-8B-Base, Cost=0%2026.05 | 36.9 | |
| BaseBackbone=Qwen3-4B-Base, Cost=0%2026.05 | 31.6 | |
| RLVRBackbone=Qwen2.5-Math-1.5B, Cost=100%2026.05 | 31.5 | |
| RELEXBackbone=Qwen2.5-Math-1.5B, Cost=15%2026.05 | 31.4 | |
| Weight Extrap.Backbone=Qwen2.5-Math-1.5B, Cost=15%2026.05 | 30.6 | |
| ExPOBackbone=Qwen2.5-Math-1.5B, Cost=15%2026.05 | 29.5 | |
| AlphaRLBackbone=Qwen2.5-Math-1.5B, Cost=15%2026.05 | 28.4 | |
| Logits Extrap.Backbone=Qwen2.5-Math-1.5B, Cost=15%2026.05 | 28.2 | |
| BaseBackbone=Qwen2.5-Math-1.5B, Cost=0%2026.05 | 23.2 |