Visual Question Answering on V*Bench Original
91.6AccuracySeed-2.0-Lite
Evaluation results
| Method | Links | |
|---|---|---|
| Seed-2.0-Lite2026.06 | 91.6 | |
| Qwen3.6-Plus2026.06 | 90.6 | |
| Gemini 3.1 Pro2026.06 | 88 | |
| GPT-5.52026.06 | 86.9 | |
| Kimi K2.62026.06 | 81.7 | |
| Claude Opus 4.72026.06 | 78.5 | |
| MiMo v2.52026.06 | 78 | |
| Llama 4 Maverick2026.06 | 61.3 |