| agents-on-agents-last-exam-ale-cli | DeepSeek-V4.1-Flash-1017 | #1 | Pass Rate: 35.4 |
| agents-on-agents-last-exam-ale-cli | DeepSeek-V4.1-Flash Max | #2 | Pass Rate: 31.8 |
| agents-on-automationbench-public | DeepSeek-V4.1-Flash-1017 | #1 | Success Rate: 60.5 |
| agents-on-automationbench-public | DeepSeek-V4.1-Flash Max | #2 | Success Rate: 54.8 |
| code-generation-on-nl2repo | DeepSeek-V4.1-Flash-1017 | #1 | Accuracy: 68.2 |
| code-generation-on-nl2repo | DeepSeek-V4.1-Flash Max | #2 | Accuracy: 65.4 |
| code-generation-on-terminal-bench-2-1 | DeepSeek-V4.1-Flash Max (DeepSeek Harness (Minimal)) | #1 | Accuracy: 90.6 |
| mathematical-reasoning-on-aime-2026 | DeepSeek-V4.1-Flash Max | #1 | Accuracy: 100.0 |
| vulnerability-reproduction-on-cybergym | DeepSeek-V4.1-Flash Max (DeepSeek Harness (Minimal)) | #1 | Success Rate: 88.1 |
| code-generation-on-deepswe | DeepSeek-V4.1-Flash-1017 (Mini-SWE-V2) | #2 | Pass@1: 0.77 |
| code-generation-on-deepswe | DeepSeek-V4.1-Flash Max (mini-SWE) | #3 | Pass@1: 0.742 |
| chart-question-answering-on-chartography | DeepSeek-V4.1-Flash-1017 (w/ tools) | #3 | Pass@1: 0.83 |
| chart-question-answering-on-chartography | DeepSeek-V4.1-Flash (with tools, max effort) | #4 | Pass@1: 0.789 |
| code-generation-on-terminal-bench-3-0 | DeepSeek-V4.1-Flash Max (DeepSeek Harness (Minimal)) | #4 | Accuracy: 30.0 |
| multi-task-language-understanding-on-mgsm | DeepSeek-V4.1-Flash Base | #5 | Average (%): 80.2 |
| code-generation-on-terminal-bench-4-0 | DeepSeek-V4.1-Flash-1017 (Claude Code) | #6 | Accuracy: 42.3 |
| code-generation-on-terminal-bench-4-0 | DeepSeek-V4.1-Flash Max (DeepSeek Harness (Minimal)) | #8 | Accuracy: 31.2 |
| multiple-choice-question-answering-mcqa-on-gpqa-diamond | DeepSeek-V4.1-Flash Max | #9 | Accuracy: 90.9 |
| code-generation-on-programbench | DeepSeek-V4.1-Flash Max (DeepSeek Harness (Minimal)) | #12 | Almost Resolved: 20.3 |
| code-generation-on-programbench | DeepSeek-V4.1-Flash-1017 | #13 | Almost Resolved: 29.1 |
| long-context-understanding-on-longbench-v2 | DeepSeek-V4.1-Flash Base | #16 | Accuracy: 45.2 |
| sentence-completion-on-hellaswag | DeepSeek-V4.1-Flash Base | #25 | Accuracy: 87.2 |
| arithmetic-reasoning-on-gsm8k | DeepSeek-V4.1-Flash Base | #43 | Accuracy: 93.0 |
| math-word-problem-solving-on-math | DeepSeek-V4.1-Flash Base | #49 | Accuracy: 61.1 |
| multi-task-language-understanding-on-mmlu-pro | DeepSeek-V4.1-Flash Base | #60 | Accuracy: 74.1 |
| code-generation-on-humaneval | DeepSeek-V4.1-Flash Base | #62 | Pass@1: 79.4 |