Vision on CVBench
81.5CVBench ScoreFLARE-X 8B
Evaluation results
| Method | Links | |
|---|---|---|
| FLARE-X 8B# Vis tok.=14002025.04 | 81.5 | |
| Qwen2.5VL 7B# Vis tok.=14002025.04 | 81.1 | |
| FLARE-X 3B# Vis tok.=14002025.04 | 80.1 | |
| FLARE-L 8B# Vis tok.=6302025.04 | 79.5 | |
| FLARE-L 3B# Vis tok.=6302025.04 | 78.2 | |
| SigLIP2 | Qwen-3-8BVision Encoder=SigLIP2, LLM=Qwen-3-8B, Model Adaptation Type=Internal E-commerce Adaptation2026.02 | 77.2 | |
| Qwen2.5VL 3B# Vis tok.=14002025.04 | 75.5 | |
| SigLIP2 | Qwen-3-4BVision Encoder=SigLIP2, LLM=Qwen-3-4B, Model Adaptation Type=Internal E-commerce Adaptation2026.02 | 73.7 | |
| Florence-VL 8B# Vis tok.=5762025.04 | 73.4 | |
| SigLIP2 | e-Llama3.1-8BVision Encoder=SigLIP2, LLM=e-Llama3.1-8B, Model Adaptation Type=Internal E-commerce Adaptation2026.02 | 72.7 | |
| Cambrain 8B# Vis tok.=5762025.04 | 72.2 | |
| SigLIP2 | Lilium-8BVision Encoder=SigLIP2, LLM=Lilium-8B, Model Adaptation Type=Internal E-commerce Adaptation2026.02 | 71.4 | |
| Florence-VL 3B# Vis tok.=5762025.04 | 70.2 | |
| SigLIP | Gemma3-4BVision Encoder=SigLIP, LLM=Gemma3-4B, Model Adaptation Type=Internal E-commerce Adaptation2026.02 | 69.8 | |
| Phi 3.5-Vision# Vis tok.=21002025.04 | 69.3 | |
| SigLIP2 | Llama-3.1-8BVision Encoder=SigLIP2, LLM=Llama-3.1-8B, Model Adaptation Type=Internal E-commerce Adaptation2026.02 | 62.3 | |
| SigLIP2 | Lilium-4BVision Encoder=SigLIP2, LLM=Lilium-4B, Model Adaptation Type=Internal E-commerce Adaptation2026.02 | 61.8 | |
| Qwen2.5ViT | e-Llama3.1-8BVision Encoder=Qwen2.5ViT, LLM=e-Llama3.1-8B, Model Adaptation Type=Internal E-commerce Adaptation2026.02 | 61.6 | |
| SigLIP2 | Lilium-1BVision Encoder=SigLIP2, LLM=Lilium-1B, Model Adaptation Type=Internal E-commerce Adaptation2026.02 | 57.4 | |
| Qwen2.5-VLVision Encoder=Qwen2.5ViT, LLM=Qwen2-7B, Model Adaptation Type=Open Source2026.02 | 32.8 | |
| LLaVA-OVVision Encoder=SigLIP, LLM=Qwen2-7B, Model Adaptation Type=Open Source2026.02 | 27.9 | |
| Qwen3-VLVision Encoder=Qwen3ViT, LLM=Qwen3-8B, Model Adaptation Type=Open Source2026.02 | 26.6 | |
| Gemma3Vision Encoder=SigLIP, LLM=Gemma3-4B, Model Adaptation Type=Open Source2026.02 | 11.4 |