| mathematical-reasoning-on-hmmt-feb-2025 | Step 3.5 Flash (11B) | #1 | Accuracy: 98.4 |
| agents-on-bench | Step-3.5-Flash | #2 | Accuracy: 88.2 |
| mathematical-reasoning-on-aime-2025 | Step-3.5-Flash | #2 | Accuracy: 97.3 |
| instruction-following-on-multichallenge | Step-3.5-Flash PaCoRe | #5 | Accuracy: 60.8 |
| instruction-following-on-multichallenge | Step-3.5-Flash | #9 | Accuracy: 55.7 |
| multi-task-language-understanding-on-mmlu | Step-3.5-Flash Base | #7 | Average (%): 85.8 |
| long-context-understanding-on-longbench-v2 | Step-3.5-Flash | #10 | Accuracy: 57.5 |
| question-answering-on-humanity-s-last-exam-hle | Step-3.5-Flash PaCoRe | #16 | Accuracy: 27.9 |
| question-answering-on-humanity-s-last-exam-hle | Step-3.5-Flash | #22 | Accuracy: 23.1 |
| sentence-completion-on-hellaswag | Step-3.5-Flash Base | #16 | Accuracy: 90.2 |
| multi-task-language-understanding-on-mmlu-pro | Step-3.5-Flash PaCoRe | #18 | Accuracy: 84.8 |
| multi-task-language-understanding-on-mmlu-pro | Step-3.5-Flash | #20 | Accuracy: 84.4 |
| multi-task-language-understanding-on-mmlu-pro | Step-3.5-Flash Base | #90 | Accuracy: 62.3 |
| instruction-following-on-ifbench | Step-3.5-Flash | #22 | Accuracy: 67.4 |
| instruction-following-on-ifbench | Step-3.5-Flash PaCoRe | #32 | Accuracy: 56.8 |
| code-generation-on-swe-bench-multilingual | Step-3.5-Flash | #25 | Accuracy: 67.4 |
| code-generation-on-swe-bench-verified | Step-3.5-Flash | #26 | Accuracy: 74.4 |
| deep-research-agents-on-browsecomp | Step-3.5-Flash (w/ context manager) | #29 | Accuracy: 69.0 |
| deep-research-agents-on-browsecomp | Step-3.5-Flash | #41 | Accuracy: 51.6 |
| code-generation-on-mbpp | Step-3.5-Flash Base | #36 | Accuracy: 79.4 |
| common-sense-reasoning-on-winogrande | Step-3.5-Flash Base | #36 | Accuracy: 79.1 |
| multiple-choice-question-answering-mcqa-on-gpqa-diamond | Step-3.5-Flash PaCoRe | #38 | Accuracy: 85.0 |
| multiple-choice-question-answering-mcqa-on-gpqa-diamond | Step-3.5-Flash | #44 | Accuracy: 83.5 |
| math-word-problem-solving-on-math | Step-3.5-Flash Base | #40 | Accuracy: 66.8 |
| code-generation-on-humaneval | Step-3.5-Flash Base | #56 | Pass@1: 81.1 |
| arithmetic-reasoning-on-gsm8k | Step-3.5-Flash Base | #90 | Accuracy: 88.2 |