Cambrian-S Release Image - MME Perception: 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. MME perception native sum of per-category accuracy and accuracy-plus scores; range0–2000. Comma is a grouping separator, not a decimal separator. Metric: MME perception score. Higher is better. Source: https://arxiv.org/html/2511.04670v1
Metric: MME perception score. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Mini-Gemini-HD-8B | 1606 | |
| 2 | LLaVA-NeXT-8B | 1603.7 | |
| 3 | Cambrian-S-7B | 1598.4 | |
| 4 | Cambrian-1-8B | 1547.1 | |
| 5 | Cambrian-S-3B | 1485.6 | |
| 6 | Cambrian-S-1.5B | 1394.4 | |
| 7 | Cambrian-S-0.5B | 1190.8 |
Interactive version: theaggregate.ai/benchmark?slug=cambrian-s-release-image-mme-perception · How It Works · Data refreshed daily, snapshot 2026-10-17.