HELM Robo-Reward-Bench - Utokyo Pr2 Opening Fridge Converted Externally To Rlds: leaderboard
The "Utokyo Pr2 Opening Fridge Converted Externally To Rlds" subtask of HELM Robo-Reward-Bench. Stanford HELM's Robo-Reward-Bench: vision-language models act as reward models over robot manipulation episodes, predicting task progress or success from video.
Metric: Absolute error. Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 0.71 | #155 |
| 2 | GPT-5 | 0.75 | #101 |
| 3 | O1 (2024-12-17) | 0.86 | #153 |
| 4 | Gemini 2.5 Flash | 0.88 | #247 |
| 5 | GPT-5 Mini (2025-08-07) | 1.02 | #175 |
| 6 | Gemini 2.0 Flash | 1.04 | #345 |
| 7 | GPT-4.1 (2025-04-14) | 1.09 | #252 |
| 8 | GPT-5 Nano | 1.11 | #430 |
| 9 | GPT-4o (2024-11-20) | 1.11 | #379 |
| 10 | Gemini 2.0 Flash Lite | 1.2 | #455 |
| 11 | GPT-4.1 Mini | 1.25 | #358 |
| 12 | Qwen2.5-Omni-7B | 1.59 | #634 |
| 13 | Gemini 2.5 Flash Lite | 1.62 | #426 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-robo-reward-bench-utokyo-pr2-opening-fridge-converted-externally-to-rlds · How It Works · Data refreshed daily, snapshot 2026-10-24.