返回
原文
Rohan Paul@rohanpaul_ai
34AI 编辑部评分，满分 100

Kimi K3 写 H100 CUDA 内核速度比优化 PyTorch 快 14.82 倍

2026-07-19 11:13· 25天前
AI 导读

Kimi K3 在 KernelBench 测试中，生成 H100 CUDA 内核的速度比优化后的 PyTorch 快 14.82 倍，性能几乎追平 Claude Opus 4.8。KernelBench 要求 AI 智能体基于可读的 PyTorch 实现，生成既正确又更快的自定义 GPU 实现，代码必须通过编译、数值正确性验证并在实际硬件上超越基准。

Kimi K3 now writes a H100 CUDA kernel 14.82x faster than optimized PyTorch, nearly matching Claude Opus 4.8.

KernelBench tests hardware work, not just whether generated code successfully compiles.

KernelBench gives an AI agent a readable PyTorch implementation and asks it to produce a custom GPU implementation that is both correct and faster.

The code must compile, return the right numerical results across test inputs and beat a reference implementation on actual hardware. A kernel that is fast but wrong fails. A kernel that is correct but slow is not particularly useful.

Xinyu YangNow you can ask K3 to write the kernels From https://kernelbench.com
编码评测/基准

来源：Rohan Paul · x.com

导出 Markdown
返回
AI 评分 34/100在 X 查看原推
导出 Markdown

Kimi K3 写 H100 CUDA 内核速度比优化 PyTorch 快 14.82 倍

Rohan Paul · @rohanpaul_ai · X·2026-07-19 11:13·25天前
AI 导读

Kimi K3 在 KernelBench 测试中，生成 H100 CUDA 内核的速度比优化后的 PyTorch 快 14.82 倍，性能几乎追平 Claude Opus 4.8。KernelBench 要求 AI 智能体基于可读的 PyTorch 实现，生成既正确又更快的自定义 GPU 实现，代码必须通过编译、数值正确性验证并在实际硬件上超越基准。

Kimi K3 now writes a H100 CUDA kernel 14.82x faster than optimized PyTorch, nearly matching Claude Opus 4.8.

KernelBench tests hardware work, not just whether generated code successfully compiles.

KernelBench gives an AI agent a readable PyTorch implementation and asks it to produce a custom GPU implementation that is both correct and faster.

The code must compile, return the right numerical results across test inputs and beat a reference implementation on actual hardware. A kernel that is fast but wrong fails. A kernel that is correct but slow is not particularly useful.

Xinyu YangNow you can ask K3 to write the kernels From https://kernelbench.com

来源：Rohan Paul· x.com

编码评测/基准