All oneshots from Kimi-K3, looks better than opus4.8.

I've ran Kimi-k3 through 34 oneshot prompts and evaluated the generated htmls, screenshots and gifs using sonnet 4.6. It came out to be better than opus4.8 from the evals. Kimi K3: https://oneshotlm.com/model/moonshotai-kimi-k3/ Opus 4.8: https://oneshotlm.com/model/anthropic-claude-opus-4-8/

中文全文翻译正在补齐，可先查看下方英文原文。

I've ran Kimi-k3 through 34 oneshot prompts and evaluated the generated htmls, screenshots and gifs using sonnet 4.6. It came out to be better than opus4.8 from the evals. Kimi K3: https://oneshotlm.com/model/moonshotai-kimi-k3/ Opus 4.8: https://oneshotlm.com/model/anthropic-claude-opus-4-8/