Cambrian-S Release VSI-Bench - Obj. Count: leaderboard
Cambrian-S release paper; final Stage4 checkpoint. Benchmark-specific sampling of Section3 retained, dagger-marked video scores use paper protocol; no pretrained, intermediate-stage or LFP checkpoint merge. Numerical columns use source mean-relative-accuracy, multiple-choice columns use accuracy. No human or chance reference becomes a model. Metric: Mean relative accuracy (%). Higher is better. Source: https://arxiv.org/html/2511.04670v1
Metric: Mean relative accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 1.5 Pro | 56.2 | #443 |
| 2 | Gemini 1.5 Flash | 49.8 | #615 |
| 3 | GPT-4o | 46.2 | #339 |
| 4 | Gemini 2.5 Pro | 43.8 | #151 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=cambrian-s-release-vsi-bench-obj-count · How It Works · Data refreshed daily, snapshot 2026-10-17.