Artificial Analysis 测评：Claude 升档位比拉满 effort 更划算
Artificial Analysis 用 Cost per Task 数据算了一笔账：Opus 5.5 开 xhigh 比 Sonnet 5.5 拉满 effort 又快又便宜，Fable 5.2 也压过 Opus。做 Claude API 选型的人该看看这份对比，别只盯着 effort 参数调。
Artificial Analysis 对 Anthropic 当前 Claude 系列做了 Cost per Task 与 Time per Task 对比，结论是升级模型档位比在同一模型上拉满 effort 参数更能换来每美元智能。Claude Opus 5.5（xhigh，得分 57）超过 Claude Sonnet 5.5（max，得分 55），且 Cost per Task 低 40%，Time per Task 约为其一半。
进一步看，Claude Fable 5.2（high）的表现超过 Opus 5.5（max），成本和时间都有节省；Fable 5.2（low）则超过 Opus 5.5（high）。也就是说，在大模型梯队里，每一档更大的模型几乎都落在性能-成本帕累托前沿上，与其烧更多推理算力，不如直接换更大的模型。
这一规律在最低档不成立：从 Claude Haiku 5.5（max）再往上走，每一步都意味着成本上升，性价比反而下降。对开发者和企业采购来说，选型的第一杠杆是模型档位，effort 参数更适合在已定档位内做微调。
With the current Claude lineup, moving up a model tier beats maxing out effort to achieve the highest intelligence per dollar
At nearly every Cost per Task, we see the largest model sits on the performance vs. cost Pareto frontier for @AnthropicAI models:
➤ Claude Opus 5.5 (xhigh, 57) outscores Claude Sonnet 5.5 (max, 55) with 40% lower Cost per Task and about half the Time per Task ➤ Claude Fable 5.2 (high) beats Opus 5.5 (max), with smaller cost and time savings, and Fable 5.2 (low) beats Opus 5.5 (high) ➤ This pattern doesn’t hold at the cheapest tier, where every step up from Claude Haiku 5.5 (max) costs more
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Claude, Anthropic, Opus 5.5, Sonnet 5.5.
Source attribution: curated from Artificial Analysis (@ArtificialAnlys).
