Mathematics on HRM8K
89.5Scoregpt-oss-120b
Evaluation results
| Method | Links | |
|---|---|---|
| gpt-oss-120bParameter Count=117B, Configuration Variant=high2026.01 | 89.5 | |
| Solar OpenParameter Count=102B2026.01 | 87.6 | |
| Qwen3-235B-A22B Instruct-2507Architecture=MoE, Total Params=235B, Activated Params=22B2026.01 | 86.1 | |
| GLM-4.5-AirParameter Count=110B2026.01 | 86 | |
| gpt-oss-120bParameter Count=117B, Configuration Variant=medium2026.01 | 84.8 | |
| DeepSeek-V3.2Mode=Non-Reasoning, Architecture=MoE, Total Params=671B, Activated Params=37B2026.01 | 83.1 | |
| K-EXAONEMode=Non-Reasoning, Architecture=MoE, Total Params=236B, Activated Params=23B2026.01 | 81.4 | |
| EXAONE 4.0Mode=Non-Reasoning, Architecture=Dense, Total Params=32B, Activated Params=32B2026.01 | 73.7 |