gemini-robotics-er-1.5-preview ranks first on Point-Bench.
Google DeepMind Sep 2025 Closed weights Mid-size
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| GPQA Diamond | 83.3% | #91 of 284 | Model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Aider Polyglot | 57.3% | #16 of 29 | Model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Point-Bench | 67.1% | #1 of 17 | Official leaderboard | |
| MMMU | 80.7% | #16 of 97 | Model card | |
| RoboRewardBench | 0.91 | #8 of 19 | Official leaderboard | |
| VPCT | 40.8% | #12 of 28 | Epoch AI | |
| MVPBench | 72.5% | #1 of 2 | Model card | |
| Where2Place | 59% | #1 of 2 | Model card | |
| CVBench | 84.3% | #6 of 11 | Model card | |
| ClockBench | 15% | #22 of 37 | Official leaderboard | |
| ERQA | 54.8% | #18 of 26 | Model card | |
| BLINK | 57.8% | #21 of 28 | Model card | |
| EmbSpatialBench | 78.4% | #10 of 10 | Model card | |
| RefSpatial-Bench | 48.5% | #8 of 8 | Model card | |
| VSI-Bench | 45.8% | #4 of 4 | Model card | |
| MindCube | 54.7% | #3 of 3 | Model card | |
| OpenEQA | 55.0 | #2 of 2 | Model card | |
| SAT (Spatial Aptitude Training) | 76.7% | #1 of 1 | Model card |
Every result in the longer documents the lab published with it. 16 of them aren't in the launch post.