SaaS-Bench - Multimodal: leaderboard
Metric: Checkpoint score (%) on the 32 multimodal tasks (images, documents and media files as inputs), weighted-checkpoint score over the final application states, browser-use agent loop over live self-hosted SaaS apps (DOM and screenshot observations, restricted browser actions, no JavaScript or backend access), same step budget, timeout and failure cap for every model; higher is better. Source: arxiv.org. Saturation forecast: Around 2030. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 | 48.7 |
| 2 | GPT-5.5 (High) | 47.6 |
| 3 | Claude Opus 4.7 | 46.3 |
| 4 | GPT-5.4 (High) | 46.1 |
| 5 | Qwen 3.6 Plus | 45.5 |
| 6 | Seed 2.0 Pro | 44.2 |
| 7 | Kimi K2.6 | 43.5 |
| 8 | Gemini 3.1 Pro (Preview) | 42.4 |
| 9 | Kimi K2.5 | 37 |
| 10 | Gemini 3.5 Flash (High) | 36.5 |
| 11 | Claude Sonnet 4.6 | 33.9 |
Interactive version: theaggregate.ai/benchmark?slug=saas-bench-multimodal · How It Works · Data refreshed daily, snapshot 2026-10-07.