Visual Instruction Following on FronTalk Single-Turn 1.0 (test)
77.5PRClaude-4-Sonnet
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Claude-4-SonnetNumber of Parameters=-, Feedback Type=Textual2025.12 | 77.5 | 71.8 | |
| GPT-OSSNumber of Parameters=120B, Feedback Type=Textual2025.12 | 68 | 73.5 | |
| Qwen3-CoderNumber of Parameters=480B, Feedback Type=Textual2025.12 | 67 | 64 | |
| GPT-OSSNumber of Parameters=20B, Feedback Type=Textual2025.12 | 64.3 | 68.5 | |
| Qwen3Number of Parameters=235B, Feedback Type=Textual2025.12 | 62.9 | 57.8 | |
| DeepSeek-R1Number of Parameters=685B, Feedback Type=Textual2025.12 | 60 | 53.5 | |
| Gemini-2.5-ProNumber of Parameters=-, Feedback Type=Textual2025.12 | 57.2 | 56.8 | |
| Qwen3-VLNumber of Parameters=30B, Feedback Type=Textual2025.12 | 54.6 | 50 | |
| Qwen3Number of Parameters=8B, Feedback Type=Textual2025.12 | 54.3 | 45.8 | |
| Qwen3-CoderNumber of Parameters=30B, Feedback Type=Textual2025.12 | 54.1 | 49.8 | |
| Ovis2.5Number of Parameters=9B, Feedback Type=Textual2025.12 | 51.5 | 49.3 | |
| GPT-4oNumber of Parameters=-, Feedback Type=Textual2025.12 | 51.4 | 50 | |
| Gemma3Number of Parameters=12B, Feedback Type=Textual2025.12 | 47.1 | 43.5 | |
| Qwen2.5-VLNumber of Parameters=32B, Feedback Type=Textual2025.12 | 44.5 | 38 | |
| Qwen2.5-VLNumber of Parameters=72B, Feedback Type=Textual2025.12 | 43 | 33.8 | |
| Llama-3.3Number of Parameters=70B, Feedback Type=Textual2025.12 | 38.4 | 35.8 | |
| Gemma3Number of Parameters=27B, Feedback Type=Textual2025.12 | 36.7 | 37.2 | |
| GLM-4.1V-ThinkingNumber of Parameters=9B, Feedback Type=Textual2025.12 | 36.6 | 32.3 | |
| Qwen2.5-VLNumber of Parameters=7B, Feedback Type=Textual2025.12 | 24.5 | 19.5 | |
| GLM-4.5VNumber of Parameters=108B, Feedback Type=Textual2025.12 | 24.4 | 23.5 |