# AI模型性价比新格局：Fable 5、Grok 4.5与DeepSeek V4对比

- 来源：Artificial Analysis (@ArtificialAnlys)
- 发布时间：2026-07-24 04:09
- AIHOT 分数：59
- AIHOT 链接：https://aihot.virxact.com/items/cmrxymopm005jropgoxkw61fn
- 原文链接：https://x.com/ArtificialAnlys/status/2080384902571139386

## AI 摘要

Artificial Analysis发布旧金山数字广告牌，展示最新AI模型智能指数与成本对比。Anthropic的Fable 5以60分领先，但平均任务成本$2.75；xAI的Grok 4.5（高）得分54，成本仅$0.31，便宜约9倍；DeepSeek V4 Flash智能得分44，任务成本$0.04，比Fable便宜约69倍。三者均位于帕累托前沿，代表当前最优性价比模型。

## 正文

Our digital billboards across San Francisco now feature our Artificial Analysis Intelligence Index vs. Cost per Task chart, plotting the Intelligence/Cost Pareto frontier of today's leading models

Cost efficiency is becoming a primary factor in evaluating model performance. Fable 5 from @AnthropicAI leads the Index at 60 but averages $2.75 Cost per Task. Grok 4.5 (high) from @elonmusk's @SpaceXAI scores 54 on the Artificial Analysis Intelligence Index but at $0.31 is ~9x cheaper and also sits on the Pareto frontier. DeepSeek v4 Flash is again lower in intelligence at 44 but is ~69x cheaper than Fable at $0.04 Cost per Task.

The Pareto frontier is the set of models that no other model beats on both intelligence and cost at their respective positions. Everything else is paying more for less.