Progress Estimation on Behavior
87.08MRAPRIMO R1
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PRIMO R1Parameters=7B2026.03 | 87.08 | 12.9 | |
| GPT-4oAccess=Closed-Source2026.03 | 79.92 | 20.08 | |
| GPT-5 miniAccess=Closed-Source2026.03 | 79.6 | 20.08 | |
| Qwen2.5-VL-72BAccess=Open-Source, Parameters=72B2026.03 | 79.49 | 20.5 | |
| Gemini 2.5 FlashAccess=Closed-Source2026.03 | 78.01 | 21.99 | |
| Robobrain 7BType=Reasoning & Video MLLM, Parameters=7B2026.03 | 72.52 | 26.97 | |
| InternVL 3.5 8BAccess=Open-Source, Parameters=8B2026.03 | 72.19 | 27.15 | |
| Claude-Haiku-4.5Access=Closed-Source2026.03 | 70.93 | 29.07 | |
| Video R1 7BType=Reasoning & Video MLLM, Parameters=7B2026.03 | 70.63 | 29.37 | |
| Qwen2.5-VL-7BAccess=Open-Source, Parameters=7B2026.03 | 69.91 | 30.06 | |
| Cosmos-Reasoning 7BType=Reasoning & Video MLLM, Parameters=7B2026.03 | 67.06 | 32.35 | |
| ProgressLM-3B-RLType=Reasoning & Video MLLM, Parameters=3B2026.03 | 46.43 | 28.23 |