# Grok 4.5 以 13 倍低成本击败 Kimi K3

- 来源：Rohan Paul (@rohanpaul_ai)
- 发布时间：2026-08-07 07:45
- AIHOT 分数：34
- AIHOT 链接：https://aihot.virxact.com/items/cmsi6w8u112lzronki3ov3wng
- 原文链接：https://x.com/rohanpaul_ai/status/2085512712881357146

## AI 摘要

AI/ML API 用同一提示词、一次性任务对比四款模型生成 3D 玩偶的成本：Grok 4.5 仅 $0.15，GPT 5.6 Sol 为 $0.60，Qwen 3.8 Max 为 $0.67，而 Kimi K3 高达 $1.98。Kimi 思考约 19 分钟，费用是 Grok 的 13 倍。对生产级智能体而言，单次任务完成成本正变得比单 token 价格更重要。

## 正文

Grok 4.5 beat Kimi K3 on the same prompt at 13x lower cost.

Very interesting experiment by AI/ML API (@aimlapi)

Same prompt, same one-shot task

Cost per figure:
Grok 4.5 - $0.15
GPT 5.6 Sol - $0.60
Qwen 3.8 Max - $0.67
Kimi K3 - $1.98

Kimi spent ~19 minutes thinking and billed 13x more. Grok just shipped.

For production agents, cost per completed task is starting to matter a lot more than cost per token.

### 引用推文

> AI/ML API：Grok 4.5 just beat Kimi K3 at 13x cheaper price! $0.15 vs $1.98 - same task, same prompt, one shot. We'll miss Nikita Bier - the best Head of Product X ever had...