Claude Opus 4.7 Thinking is now #1 in Vision Arena. This is a +3 point improvement over Opus 4.6 Thi...
- Opus 4.7在图表、作业、OCR等视觉子任务中表现显著提升，最高单项增益达30分。
- 该模型目前是视觉输入推理能力最强的AI系统，在多个细分类别全面领先。
- 非Anthropic阵营最佳模型Muse Spark排名第四，与Opus 4.7存在13分性能差距。
This is a +3 point improvement over Opus 4.6 Thinking and +13 points over the next non-Anthropic model, Muse Spark (#4).
The model leads across reasoning over visual inputs, sweeping multiple Vision sub-categories with the https://t.co/XCK6cYaq4r" / X
Post
Conversation
Claude Opus 4.7 Thinking is now #1 in Vision Arena. This is a +3 point improvement over Opus 4.6 Thinking and +13 points over the next non-Anthropic model, Muse Spark (#4). The model leads across reasoning over visual inputs, sweeping multiple Vision sub-categories with the biggest gains over Claude Opus 4.6 in: - #1 Diagram, Thinking: 1339 (+20) - #1 Homework, Non-Thinking: 1352 (+30) - #1 OCR, Non-Thinking: 1319 (+7)
