| Benchmark | Model | Rank | Results |
|---|---|---|---|
| code-generation-on-swe-bench-verified | Macaron-V1-Venti (Claude Code) | #3 | Accuracy: 85.6 |
| code-generation-on-terminal-bench-2-1 | Macaron-V1-Venti (Claude Code) | #6 | Accuracy: 87.6 |
| optical-character-recognition-ocr-on-ocrbench | Macaron-V1-Tall | #9 | Score: 896 |
| code-generation-on-deepswe | Macaron-V1-Venti (Claude Code) | #19 | Pass@1: 0.584 |
| multimodal-reasoning-on-mmmu | Macaron-V1-Tall | #56 | Accuracy: 61.22 |