Code Question Answering on Python Code Question Answering downstream
84AccuracyQwen-3-VL
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen-3-VLInput Modality=Text2026.02 | 84 | |
| Gemini-2.5-ProInput Modality=Text2026.02 | 82.2 | |
| GPT-5-miniInput Modality=Text2026.02 | 82 | |
| GPT-5.1Input Modality=Text2026.02 | 79.3 | |
| GLM-4.6vInput Modality=Text2026.02 | 78.5 | |
| GPT-5-miniInput Modality=Image2026.02 | 77.5 | |
| Gemini-3-ProInput Modality=Image2026.02 | 77.2 | |
| Gemini-3-FlashInput Modality=Image2026.02 | 74.8 | |
| Gemini-3-ProInput Modality=Text2026.02 | 74.8 | |
| Gemini-3-FlashInput Modality=Text2026.02 | 73.4 | |
| GLM-4.6vInput Modality=Image2026.02 | 72.6 | |
| Gemini-2.5-ProInput Modality=Image2026.02 | 71.2 | |
| GPT-5.1Input Modality=Image2026.02 | 68.6 | |
| Qwen-3-VLInput Modality=Image2026.02 | 58.1 | |
| Gemini-3-FlashInput Modality=NoCtx2026.02 | 47.9 | |
| Gemini-3-ProInput Modality=NoCtx2026.02 | 46.8 | |
| GPT-5-miniInput Modality=NoCtx2026.02 | 46.3 | |
| Qwen-3-VLInput Modality=NoCtx2026.02 | 45.6 | |
| GPT-5.1Input Modality=NoCtx2026.02 | 44.7 | |
| Gemini-2.5-ProInput Modality=NoCtx2026.02 | 40.2 | |
| GLM-4.6vInput Modality=NoCtx2026.02 | 37.2 |