| code-generation-on-livecodebench-v5 | Gemini 2.5 Flash | – | Pass@1: 59.3 |
| code-generation-on-livecodebench-v5 | Gemini 2.5 Pro | – | Pass@1: 74.2 |
| code-generation-on-swe-bench-verified | Gemini 2.5 Flash (single attempt) | – | Accuracy: 48.9 |
| code-generation-on-swe-bench-verified | Gemini 2.5 Pro (single attempt) | – | Accuracy: 59.6 |
| long-context-understanding-on-mrcr-v2-128k | Gemini 2.5 Flash | – | Accuracy: 54.3 |
| long-context-understanding-on-mrcr-v2-128k | Gemini 2.5 Pro | – | Accuracy: 58.0 |
| long-context-understanding-on-mrcr-v2-1m | Gemini 2.5 Flash | – | Accuracy: 21.0 |
| long-context-understanding-on-mrcr-v2-1m | Gemini 2.5 Pro | – | Accuracy: 16.4 |
| mathematical-reasoning-on-aime-2025 | Gemini 2.5 Flash | – | Accuracy: 72.0 |
| mathematical-reasoning-on-aime-2025 | Gemini 2.5 Pro | – | Accuracy: 88.0 |
| multimodal-reasoning-on-mmmu | Gemini 2.5 Flash | – | Accuracy: 79.7 |
| multimodal-reasoning-on-mmmu | Gemini 2.5 Pro | – | Accuracy: 82.0 |
| multimodal-reasoning-on-zerobench | Gemini 2.5 Flash | – | Accuracy: 2.0 |
| multimodal-reasoning-on-zerobench | Gemini 2.5 Pro | – | Accuracy: 4.5 |
| multiple-choice-question-answering-mcqa-on-gpqa-diamond | Gemini 2.5 Flash | – | Accuracy: 82.8 |
| multiple-choice-question-answering-mcqa-on-gpqa-diamond | Gemini 2.5 Pro | – | Accuracy: 86.4 |
| question-answering-on-humanity-s-last-exam-hle | Gemini 2.5 Flash | – | Accuracy: 11.0 |
| question-answering-on-humanity-s-last-exam-hle | Gemini 2.5 Pro | – | Accuracy: 21.6 |
| question-answering-on-simpleqa | Gemini 2.5 Flash | – | Accuracy: 26.9 |
| question-answering-on-simpleqa | Gemini 2.5 Pro | – | Accuracy: 54.0 |
| video-question-answering-on-lvbench | Gemini 2.5 Flash | – | Accuracy: 62.7 |
| video-question-answering-on-lvbench | Gemini 2.5 Pro | – | Accuracy: 78.7 |
| visual-question-answering-vqa-on-vibe-eval | Gemini 2.5 Flash | – | Accuracy: 65.4 |
| visual-question-answering-vqa-on-vibe-eval | Gemini 2.5 Pro | – | Accuracy: 67.2 |