| Benchmark | Model | Rank | Results |
|---|---|---|---|
| code-generation-on-swe-bench-verified | Gemini 3 Pro (single attempt) | – | Accuracy: 76.2 |
| code-generation-on-terminal-bench-2-0-terminus-2 | Gemini 3 Pro (Terminus-2) | – | Accuracy: 54.2 |
| multimodal-reasoning-on-mmmu-pro | Gemini 3 Pro | – | Accuracy: 81 |
| multiple-choice-question-answering-mcqa-on-gpqa-diamond | Gemini 3 Deep Think | – | Accuracy: 93.8 |
| question-answering-on-humanity-s-last-exam-hle | Gemini 3 Deep Think | – | Accuracy: 41.0 |
| video-question-answering-on-video-mmmu | Gemini 3 Pro | – | Accuracy: 87.6 |