WHAT NEW IN X11.01 SUN

X 上的 AI 合集档案 · 22 个合集 · 63,170 条

benchmarkIMAGE

Nvidia 与 ChipAgents 发布芯片设计 Agent 基准

Build-A-Chip-Bench 测试 AI Agent 自主设计芯片能力,最佳 Agent 仅得 16.6%,功能验证通过率约 40%。

Choblin
@choblin2916 天前收录原推发布于 2026.10.15
Nvidia 与 ChipAgents 发布芯片设计 Agent 基准(第 1 张,共 4 张)
Nvidia 与 ChipAgents 发布芯片设计 Agent 基准(第 2 张,共 4 张)
Nvidia 与 ChipAgents 发布芯片设计 Agent 基准(第 3 张,共 4 张)
Nvidia 与 ChipAgents 发布芯片设计 Agent 基准(第 4 张,共 4 张)
Choblin @choblin29原推发布于 2026.10.15
在 X 上打开原推

🚨BREAKING: Nvidia and ChipAgents have launched Build-A-Chip-Bench, a new benchmark testing whether AI agents can autonomously design entire chips. Even the best-performing AI agent, powered by Anthropic's Claude Fable 5.1, scored just 16.6%. Across tested agents, roughly 40% of tasks passed functional verification, but only 10% met the physical requirements needed for a viable chip design. ChipAgents CEO William Wang says AI has already cut chip development timelines from at least two years to around one year. He expects further advances to bring that down to just four to six months within the next year.

x.com/choblin29/status/2110446663789842937采集于 2026.10.15 10:49