Reward Modeling on FIRM-Bench Gen
79Score (Easy)InternVL3.5-38B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| InternVL3.5-38B2026.03 | 79 | 58 | 68 | |
| GPT-4o2026.03 | 63 | 66 | 65 | |
| Qwen3-VL-8B2026.03 | 63 | 63 | 63 | |
| InternVL3.5-8B2026.03 | 61 | 55 | 58 | |
| InternVL3.5-241B-A282026.03 | 60 | 61 | 60 | |
| Qwen3-VL-235B-A222026.03 | 60 | 52 | 56 | |
| GPT-4.12026.03 | 58 | 69 | 64 | |
| Qwen3-VL-32B2026.03 | 51 | 56 | 54 | |
| GPT-52026.03 | 50 | 53 | 52 | |
| FIRM-Gen-8B2026.03 | 45 | 57 | 51 | |
| Gemini-3-pro2026.03 | 38 | 42 | 40 |