Visual Search on HR-Bench 8K (Accuracy)
77.8AccuracyInterSketch
Evaluation Results
| Method | Links | |
|---|---|---|
| InterSketch2026.05 | 77.8 | |
| Qwen2.5-VL-72BModel Category=Open-Source MLLMs2025.12 | 76.3 | |
| PyVision-ImageTooling Type=Dynamic Tooling2026.02 | 74.3 | |
| DeepEyes-v2Tooling Type=Dynamic Tooling2026.02 | 73.8 | |
| DeepEyeV2-7B2026.05 | 73.8 | |
| Mini-o3Tooling Type=Static Toolset2026.02 | 73.3 | |
| DeepEyesTooling Type=Static Toolset2026.02 | 72.6 | |
| CodeDance-7BModel Category=Open-Source MLLMs with Tools2025.12 | 72.3 | |
| CodeDanceTooling Type=Dynamic Tooling2026.02 | 72.3 | |
| CodeVisionTooling Type=Dynamic Tooling2026.02 | 72.2 | |
| Thyme-VL-7BModel Category=Open-Source MLLMs with Tools2025.12 | 72 | |
| ThymeTooling Type=Dynamic Tooling2026.02 | 72 | |
| Thyme-7B2026.05 | 72 | |
| Deepeyes-7BModel Category=Open-Source MLLMs with Tools, reproduced=true2025.12 | 71.9 | |
| Qwen2.5-VL-32BModel Category=Open-Source MLLMs2025.12 | 71.6 | |
| CodeVTooling Type=Dynamic Tooling2026.02 | 71.3 | |
| InternVL3-8BModel Category=Open-Source MLLMs2025.12 | 70 | |
| Qwen3-VL-8B2026.05 | 68.5 | |
| Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B2026.02 | 67.9 | |
| InternVL3-78BModel Category=Open-Source MLLMs2025.12 | 67.3 | |
| Pixel Reasoner-7BModel Category=Open-Source MLLMs with Tools2025.12 | 66.9 | |
| Pixel-ReasonerTooling Type=Static Toolset2026.02 | 66.9 | |
| Qwen2.5-VL-7BModel Category=Open-Source MLLMs2025.12 | 66 | |
| InternVL2.5-8BModel Category=Open-Source MLLMs2025.12 | 65.5 | |
| Qwen2.5-VL-7B2026.05 | 65.3 | |
| Llava-OneVision-72BModel Category=Open-Source MLLMs2025.12 | 60.9 | |
| Llava-OneVision-7BModel Category=Open-Source MLLMs2025.12 | 60 | |
| GPT-4oModel Category=Closed-Source MLLMs2025.12 | 59.6 | |
| ThinkMorph-7B2026.05 | 46.7 |