UniClawBench - Multimodal: leaderboard
UniClawBench (HKU MMLab, 2026) runs proactive agents on 400 bilingual real-world tasks in live Docker containers under the OpenClaw framework, with up to two simulated user follow-ups and a hidden GPT-5.4 supervisor that grades step checkpoints. This board is the pass rate on the 80 multimodal tasks, in percent; higher is better.
Metric: Pass rate (%) on the 80 multimodal tasks of 400 bilingual real-world tasks run live in Docker containers under the OpenClaw framework (v2026.3.11), with up to two simulated user follow-ups and a hidden GPT-5.4 Codex supervisor that grades step checkpoints. Source: arxiv.org. Saturation forecast: Around October 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 21.2 | #161 |
| 2 | GPT-5.4 | 17.5 | #123 |
| 3 | Claude Opus 4.8 | 17.5 | #56 |
| 4 | Gemini 3.1 Pro (Preview) | 16.3 | #94 |
| 5 | Gemini 3 Flash | 8.7 | #164 |
| 6 | GPT-4.1 | 7.5 | #495 |
| 7 | Kimi K2.6 | 7.5 | #178 |
| 8 | Qwen 3.5 Plus | 7.5 | #232 |
| 9 | GPT-5.4 Mini | 6.2 | #389 |
| 10 | Gemini 3.1 Flash Lite | 3.7 | #364 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=uniclawbench-multimodal · How It Works · Data refreshed daily, snapshot 2026-11-13.