Open-world image classification on Open-world Fine-grained set
62.7True Identification (TI)SpeciaRL-7B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SpeciaRL-7BMethod Category=Reasoning LMMs - fine-tuned out-of-distribution2026.03 | 62.7 | 74.4 | 49.2 | 81.1 | |
| Qwen2.5VL-7B (rft)Method Category=Reasoning LMMs - fine-tuned out-of-distribution2026.03 | 62 | 74.8 | 48.4 | 80.6 | |
| Qwen2.5VL-7B (sft)Method Category=Reasoning LMMs - fine-tuned out-of-distribution2026.03 | 60 | 73.8 | 47.8 | 80.1 | |
| QWEN2.5VL 7BMethod Category=Reasoning LMMs2026.03 | 58.7 | 74.2 | 47 | 78.9 | |
| QWEN2.5VL 3BMethod Category=Reasoning LMMs2026.03 | 44.3 | 63.9 | 41.6 | 69.3 | |
| QWEN2VL 2BMethod Category=Non-reasoning LMMs2026.03 | 35.7 | 62.5 | 40.7 | 63.4 | |
| QWEN2VL 7BMethod Category=Non-reasoning LMMs2026.03 | 34.6 | 64 | 39.2 | 62.9 | |
| CLIP retrievalMethod Category=Retrieval-based baselines2026.03 | 32.4 | 45.4 | 42.9 | 65.4 | |
| CASEDMethod Category=Retrieval-based baselines2026.03 | 27.4 | 46.6 | 60.7 | 61.7 | |
| LLAVA-NEXT Mist 7BMethod Category=Non-reasoning LMMs2026.03 | 26.8 | 43.7 | 35.3 | 60.1 | |
| INTERNVL2 8BMethod Category=Non-reasoning LMMs2026.03 | 22.3 | 46.7 | 34.8 | 56.7 | |
| INTERNVL2.5 8BMethod Category=Reasoning LMMs2026.03 | 17.9 | 50.9 | 32.8 | 53.5 | |
| INTERNVL2.5 4BMethod Category=Reasoning LMMs2026.03 | 17.2 | 48.2 | 32.8 | 52.3 | |
| LLAVA-NEXT Vic 7BMethod Category=Non-reasoning LMMs2026.03 | 16.9 | 44.5 | 32.2 | 53.2 | |
| INTERNVL2 4BMethod Category=Non-reasoning LMMs2026.03 | 16.2 | 44.4 | 32 | 52 | |
| INTERNVL2 2BMethod Category=Non-reasoning LMMs2026.03 | 14.9 | 47 | 31.6 | 50.7 | |
| PHI-3-VISIONMethod Category=Non-reasoning LMMs2026.03 | 13.4 | 49.1 | 31.8 | 47.2 | |
| INTERNVL2.5 2BMethod Category=Reasoning LMMs2026.03 | 12.2 | 38.6 | 27.5 | 47 | |
| INSTRUCTBLIP Vic 7BMethod Category=Non-reasoning LMMs2026.03 | 10.4 | 48.8 | 35.6 | 47.2 | |
| LLAVA-1.5 7BMethod Category=Non-reasoning LMMs2026.03 | 8.4 | 46.5 | 28.2 | 44.8 | |
| LLAVA-OV Qwen2 7BMethod Category=Non-reasoning LMMs2026.03 | 6.4 | 40.4 | 39 | 43.8 | |
| LLAVA-OV Qwen2 0.5BMethod Category=Non-reasoning LMMs2026.03 | 6 | 42.7 | 38.5 | 43.3 | |
| IDEFICS2 8BMethod Category=Non-reasoning LMMs2026.03 | 3 | 49.9 | 38 | 41.7 |