(无标题)
The massive jump we've seen on ARC-AGI-2 has been a surprise to me. Now performance increased from 17.9% to 52.9%; at the same time we're seeing dramatically improved tool calling performance. https://t.co/SqhwH0dxbx
动态
The massive jump we've seen on ARC-AGI-2 has been a surprise to me. Now performance increased from 17.9% to 52.9%; at the same time we're seeing dramatically improved tool calling performance. https://t.co/SqhwH0dxbx